Abstract

Neuronal oscillations are a ubiquitous phenomenon in the human nervous system. Alpha-band oscillations (8–12 Hz) have been shown to correlate negatively with attention and performance, whereas gamma-band oscillations (40–150 Hz) correlate positively. Here, we studied the relation between prestimulus alpha-band power and poststimulus gamma-band power in a suprathreshold tactile discrimination task. Participants received two electrical stimuli to their left index finger with different SOAs (0 msec, 100 msec, intermediate SOA, intermediate SOA ± 10 msec). The intermediate SOA was individually determined so that stimulation was bistable, and participants perceived one stimulus in half of the trials and two stimuli in the other half. We measured neuronal activity with magnetoencephalography (MEG). In trials with intermediate SOAs, behavioral performance correlated inversely with prestimulus alpha-band power but did not correlate with poststimulus gamma-band power. Poststimulus gamma-band power was high in trials with low and high prestimulus alpha-band power and low for intermediate prestimulus alpha-band power (i.e., U-shaped). We suggest that prestimulus alpha activity modulates poststimulus gamma activity and subsequent perception: (1) low prestimulus alpha-band power leads to high poststimulus gamma-band power, biasing perception such that two stimuli were perceived; (2) intermediate prestimulus alpha-band power leads to low gamma-band power (interpreted as inefficient stimulus processing), consequently, perception was not biased in either direction; and (3) high prestimulus alpha-band power leads to high poststimulus gamma-band power, biasing perception such that only one stimulus was perceived.

INTRODUCTION

Even in the absence of external sensory input, the brain is constantly active. Thus, neuronal activity is constantly fluctuating (Buzsáki & Draguhn, 2004). Incoming stimuli can therefore impinge on different levels of neuronal activity (i.e., brain states) at different times. These brain states can influence the processing of stimuli (Iemi, Chaumon, Crouzet, & Busch, 2017; Lange, Keil, Schnitzler, van Dijk, & Weisz, 2014; Weisz et al., 2014; Keil, Müller, Ihssen, & Weisz, 2012; Jensen & Mazaheri, 2010).

One prominent marker of brain states is neuronal oscillation. Neuronal oscillations refer to rhythmic changes in activity of neuronal populations (Buzsáki & Watson, 2012). Thus, fluctuations of brain states can be reflected in fluctuations of these neuronal oscillations. Two prominent frequency bands are the alpha (8–12 Hz) and gamma band (40–150 Hz). It has been found that fluctuations in prestimulus alpha-band power correlate with varying perception despite physically identical stimulation (Lange, Halacz, van Dijk, Kahlbrock, & Schnitzler, 2012; van Dijk, Schoffelen, Oostenveld, & Jensen, 2008; Linkenkaer-Hansen, Nikulin, Palva, Ilmoniemi, & Palva, 2004). For example, lower parieto-occipital alpha-band power increased participants' ability to detect near-threshold visual stimuli (van Dijk et al., 2008; Hanslmayr et al., 2007). Similarly, prestimulus alpha-band power in contralateral somatosensory-posterior areas was lower when participants could discriminate veridically between two subsequent tactile stimuli compared with trials where participants perceived stimulation as one single stimulus (Baumgarten, Schnitzler, & Lange, 2016). Given these results, it was suggested that prestimulus alpha oscillations reflect the excitability of a brain area, which in turn influences the neuronal processing and perception of ambiguous stimuli (Lange et al., 2014; Lange, Oostenveld, & Fries, 2013; Thut, Nietzel, Brandt, & Pascual-Leone, 2006). In addition, alpha-band power has been related to active inhibition of brain areas (Jensen & Mazaheri, 2010; Klimesch, Sauseng, & Hanslmayr, 2007). In line with the inhibition hypothesis, prestimulus alpha-band power is modulated by spatial attention, and such modulations of alpha-band power have been shown to affect perception (Thut et al., 2006; Worden, Foxe, Wang, & Simpson, 2000; Foxe, Simpson, & Ahlfors, 1998). In addition to prestimulus alpha-band power, the power of poststimulus gamma oscillations is also modulated by attention. In visuospatial attention tasks, poststimulus gamma-band power increases in the visual area contralateral to the stimulus (e.g., Händel, Haarmeier, & Jensen, 2011; Fries, Womelsdorf, Oostenveld, & Desimone, 2008; Siegel, Donner, Oostenveld, Fries, & Engel, 2008; Müller, Gruber, & Keil, 2000). Similarly, poststimulus gamma power in tactile spatial attention tasks increases in somatosensory areas contralateral to the attended side and can affect perception (Haegens, Nácher, Hernández, et al., 2011; Haegens, Osipova, Oostenveld, & Jensen, 2010; Bauer, Oostenveld, Peeters, & Fries, 2006). Finally, it was found that poststimulus gamma oscillations and behavioral performance are linked. For example, high gamma-band power in visual cortex relates to faster RTs (Hoogenboom, Schoffelen, Oostenveld, & Fries, 2010; Womelsdorf, Fries, Mitra, & Desimone, 2006). In the somatosensory domain, higher poststimulus gamma-band power in contralateral primary somatosensory cortex (S1) relates to increased stimulus detection (Siegle, Pritchett, & Moore, 2014; Meador, Ray, Echauz, Loring, & Vachtsevanos, 2002). Generally, gamma oscillations are discussed as the neuronal underpinnings of cortical information processing (Fries, 2005, 2009, 2015).

In summary, both prestimulus alpha and poststimulus gamma oscillations are associated with attention, neuronal processing, and behavioral performance. Prestimulus alpha-band power typically decreases with higher attention, and low alpha-band power is associated with higher behavioral performance. By contrast, poststimulus gamma-band power typically increases with higher attention and high gamma-band power is associated with higher behavioral performance. Given these similar, but also diametrical effects of prestimulus alpha-band power and poststimulus gamma-band power, we speculated that prestimulus alpha-band power and poststimulus gamma-band power are directly (negatively) correlated.

To this end, we studied the relation of prestimulus alpha-band power, poststimulus gamma-band power, and tactile perception in a suprathreshold tactile discrimination task. We hypothesized that poststimulus gamma-band power in primary somatosensory cortex (S1) is positively correlated with perception, whereas prestimulus alpha-band power is negatively correlated with perception. Consequently, when comparing alpha- and gamma-band power directly, we hypothesized to find a negative correlation between prestimulus alpha-band power and poststimulus gamma-band power.

METHODS

We used data recorded by Baumgarten et al. (2016). Here, we give a concise description. More details on paradigm, participants and recordings can be found in Baumgarten et al. (2016).

Participants

We included 12 of the 16 right-handed participants (four men, mean = 26.0 years, SD = 5.3 years) measured by Baumgarten et al. (2016; see below for reasons for excluding four participants). Participants gave written informed consent in accordance with the Declaration of Helsinki and the Ethical Committee of the Medical Faculty, Heinrich-Heine-University Düsseldorf before participating in the experiment.

Participants had no known neurological disorders, no somatosensory deficits, and normal or corrected-to-normal vision.

Paradigm

Each trial began with a fixation dot in the center of the participant's visual field projected on the backside of a translucent screen (60 Hz refresh rate) positioned 60 cm in front of the participant. After 500 msec, this fixation dot decreased in luminance, indicating that the stimulation is about to be applied after a jittered period (900–1100 msec). Then, participants received two electrical stimuli (duration: 0.3 msec each) with different SOAs. Electrical stimuli were applied by electrodes located between the two distal joints of the left index finger. The amplitude of the pulses was individually determined so that stimulation was clearly perceived, but without being painful (stimulus amplitude: mean = 4.1, SD = 1.4 mA). In a premeasurement, the individual SOA was determined for which a participant veridically perceived two stimuli in ∼50% of the trials (intermediate SOA, mean = 24.6 msec, SD = 6.2 msec). During the task, participants received stimulation with five different SOAs: 0 msec, 100 msec, intermediate SOA, intermediate SOA ± 10 msec. After stimulation, the fixation dot remained visible for another jittered period (500–1200 msec) to minimize motor preparation effects. By written instruction on the screen, participants were asked to report the number of perceived stimuli (either one or two) within 3000 msec via button press with the right index or middle finger. Again, to minimize motor preparation effects, configuration of the response buttons was randomized for each trial.

Each SOA was used in 50 trials. Only the intermediate SOA was used in 200 trials, resulting in 400 trials in total. Stimuli were presented in blocks. Each block consisted of 80 trials: 40 trials with intermediate SOA and 10 trials for each of the remaining SOAs. After each block, a self-paced break (∼2 min) was included.

To familiarize participants with the task, a 5-min training phase with all five SOAs preceded the actual measurement. Before the measurement, participants received information about the task, but not about the purpose of the study or the different SOAs.

Presentation of the stimuli was done with Presentation software (Neurobehavioral Systems, Albany, NY).

Magnetoencephalography Measurement

A 306-channel whole-head magnetoencephalography (MEG; Neuromag Elekta Oy, Helsinki, Finland) was used to record brain activity at a sampling rate of 1000 Hz while participants performed the task. The MEG consisted of 102 pairs of orthogonal gradiometers and 102 magnetometers. For the analysis, only the gradiometers were taken into account. EOGs were measured to detect eye movements. EOG electrodes were placed at the outer sides of both eyes and above and below the left eye.

Data Preprocessing

Data were analyzed with custom-made scripts using Fieldtrip (Oostenveld, Fries, Maris, & Schoffelen, 2011) and Matlab (The MathWorks, Natick, MA).

Continuously recorded data were divided into trials. A trial started with the appearance of the fixation dot and ended with the press of the response button. The total number of trials was 400 with an average trial length of ∼6 sec (4–8.6 sec). Power line noise at 50 Hz and its harmonics at 100 and 150 Hz were removed by a band-stop filter, and data were bandpass filtered between 2 and 250 Hz. For the filters, we used the default options implemented in FieldTrip, that is, we used an infinite impulse response zero-phase Butterworth filter of fourth order. A mean of 5.1 (SEM = 0.5) noisy channels were removed and reconstructed by interpolation of neighboring channels. Artifacts (muscle or eye movement, SQUID jumps) were removed semiautomatically by means of a z-score-based algorithm implemented in FieldTrip, followed by an additional visual inspection to remove artifacts (e.g., extensively noisy channels or channels still containing nondetected squid jumps, etc.). A mean of 104.1 (SEM = 9.1) trials were removed due to artifacts.

Other preprocessing steps were conducted according the respective analyses (see below).

Overview of Analysis Steps

We aimed to analyze the relation between prestimulus alpha-band power, poststimulus gamma-band power, and perception. Details on the analyses will be provided below. Here, we give a concise overview of the analysis steps performed. First, for each single trial prestimulus alpha-band power was determined by averaging power in a priori defined sensors, time range, and frequency band based on results of our previous study (Baumgarten et al., 2016). Second, for each single trial poststimulus gamma-band power was determined similarly by averaging power across sensors, time, and frequency. Here, sensors of interest were determined based on the topography of the M50, and frequency ranges were determined individually.

After performing these two steps, we could determine per participant and for each single trial one value for prestimulus alpha-band power, poststimulus gamma-band power, and perception, respectively. This enabled us to sort individual trials with respect to alpha-band power or gamma-band power. Then, we combined trials to bins, computed mean gamma-band power and/or mean perception in these bins. Finally, we tested by means of first- and second-order regression analyses a putative relation between the two variables (i.e., alpha- or gamma-band power, respectively, on the one side, and gamma-band power or perception, respectively, on the other side).

Time–Frequency Analysis

Time–frequency analysis (TFA) was performed for frequencies in the alpha (8–12 Hz) and gamma band (40–150 Hz) by means of discrete Fourier transformation on sliding time windows. For the following analyses, we only used trials with intermediate SOA. Before TFA, we removed the mean of the respective time period and the linear trend. We combined each pair of gradiometers by summing the spectral power of orthogonal gradiometers. The TFA was performed on 3000-msec data segments (−1000 to 2000 msec). If the data in a trial were shorter than 3000 msec (e.g., due to removed artifacts), the corresponding trial was zero-padded to 3000 msec.

The alpha-band (8–12 Hz) power was analyzed in steps of 1 Hz with a time window Δt of seven cycles of the respective frequency ft = 7/f), moved in steps of 50 msec (Baumgarten et al., 2016). We used a single Hanning taper on each time window, resulting in spectral smoothing of 1/Δt.

In our previous study, we found a significant effect of prestimulus alpha-band power on perception in a specific set of sensors and in the prestimulus time period (−0.9 to −0.25 sec, with 0 msec being the time point in which the first electrical stimulus occurred; Baumgarten et al., 2016). Here, we thus analyzed alpha-band power in the same sensors and the same time period. As in Baumgarten et al. (2016), we averaged alpha-band power from 8 to 12 Hz in this time window and in these sensors. These sensors are as follows: MEG1042+1043, MEG1112+1113, MEG1122+1123, MEG1312+1313, MEG0712+0713, MEG0722+0723, MEG1142+1143, MEG1132+1133, MEG1342+1343, MEG2212+2213, MEG2412+2413, MEG2422+2423, MEG2642+2643, MEG1832+1833, MEG2242+2243, MEG2232+2233, MEG2012+2013, MEG2442+2443, MEG2432+2433, MEG2522+2523, MEG2312+2313, MEG2322+2323, MEG2512+2513, MEG2342+2343, MEG2022+2023, MEG2212+2213, MEG2612+2613, MEG2222+2223.

The gamma band (40–150 Hz) was analyzed in steps of 5 Hz with a time window of 100 msec, moved in steps of 20 msec. Here, we used three Slepian tapers on each time window, resulting in spectral smoothing of ±20 Hz. We focused our analysis of gamma-band power on the right primary somatosensory cortex (S1 contralateral to stimulation site) by identifying five sensors showing maximum amplitude of the M50 (MEG1122+1123, MEG1132+1133, MEG1312+1313, MEG1342+1343, MEG1332+1333; see below for details on sensor selection). In the following analyses, we averaged gamma-band power over these five sensors. Furthermore, we only used trials with intermediate SOAs.

For the analysis of gamma-band power, we first determined individual frequencies showing maximal power. To this end, we calculated for each participant, for each time point between 0 and 200 msec, and for each frequency between 40 and 150 Hz the power relative to an averaged prestimulus baseline (−600 to −200 msec) by means of an independent t test.

Next, we averaged for each frequency the t values across all poststimulus time points (0–200 msec; Baumgarten, Schnitzler, & Lange, 2017; Cousijn et al., 2014). Individual gamma-band peaks were identified using Matlab's built-in function findpeaks (Baumgarten et al., 2017). Gamma ranges with maximum power were determined by taking the width of the gamma-band peak at its half height (as implemented in the function findpeaks; Figure 1A).

Figure 1. 

Poststimulus gamma-band activity. (A) Individual spectra in the gamma-band range (40–150 Hz). Spectra were determined by computing for each frequency (40–150 Hz) and time point (0–200 msec) t values (poststimulus vs. prestimulus activity) and then averaging t values across 0–200 msec. Peaks of each spectrum were determined using the Matlab function findpeaks. Dashed horizontal lines indicate the threshold (t = 1) for a peak to be recognized. Instead of peak frequencies, our analysis relied on narrow-band frequency ranges. Frequency ranges were determined by computing the width of the peak at its half height. Smaller gray lines indicate the relative height of the peak (Prominence in Matlab function findpeaks) and the width (Width at half prominence in Matlab function findpeaks). Red vertical lines indicate the frequencies at the half height, which determine the upper and lower limits of the gamma-band range used for subsequent analyses. Note that Participants 5, 13, and 15 had to be excluded from further analyses because their gamma peaks were below the threshold. Participant 8 had to be excluded from further analyses, because increased activity extended also to lower frequencies (not shown) so that we could not excluded that this activity was actually a broadband response to stimulation. (B) Topographical representation of gamma-band activity averaged across participants. For each participant, t values in the individual gamma-band ranges (see A) were averaged for each sensor. Next, the t values were averaged across participants. Black dots indicate the sensors of interest for gamma-band analysis, which were determined beforehand.

Figure 1. 

Poststimulus gamma-band activity. (A) Individual spectra in the gamma-band range (40–150 Hz). Spectra were determined by computing for each frequency (40–150 Hz) and time point (0–200 msec) t values (poststimulus vs. prestimulus activity) and then averaging t values across 0–200 msec. Peaks of each spectrum were determined using the Matlab function findpeaks. Dashed horizontal lines indicate the threshold (t = 1) for a peak to be recognized. Instead of peak frequencies, our analysis relied on narrow-band frequency ranges. Frequency ranges were determined by computing the width of the peak at its half height. Smaller gray lines indicate the relative height of the peak (Prominence in Matlab function findpeaks) and the width (Width at half prominence in Matlab function findpeaks). Red vertical lines indicate the frequencies at the half height, which determine the upper and lower limits of the gamma-band range used for subsequent analyses. Note that Participants 5, 13, and 15 had to be excluded from further analyses because their gamma peaks were below the threshold. Participant 8 had to be excluded from further analyses, because increased activity extended also to lower frequencies (not shown) so that we could not excluded that this activity was actually a broadband response to stimulation. (B) Topographical representation of gamma-band activity averaged across participants. For each participant, t values in the individual gamma-band ranges (see A) were averaged for each sensor. Next, the t values were averaged across participants. Black dots indicate the sensors of interest for gamma-band analysis, which were determined beforehand.

We used two inclusion criteria for a frequency to be identified as a peak frequency: First, to ensure that gamma-band activity was not just a broadband signal in response to stimulation onset but a clear narrow-band range, we defined a minimum peak height relative to neighboring points (i.e., setting in findpeaks the MinPeakProminence to a t value of 0.5). By this criterion, we had to exclude one participant because we could not ensure that a seeming gamma range was actually a broadband response across a wider range of frequencies, including the beta band (20–40 Hz, Participant 8 excluded; see Figure 1A). Second, to ensure that gamma ranges with highest power were sufficiently strong to be not confused with noise fluctuations, we set an absolute threshold of t = 1 (i.e., setting in findpeaks the MinPeakHeight to a t value of 1). By this criterion, we had to exclude three participants from further analyses (Participants 5, 13, and 15; see Figure 1A).

Selection of Sensors of Interest (Event-related Field Analysis)

We focused our analysis of gamma-band power on the right primary somatosensory cortex (S1 contralateral to stimulation site). To this end, we determined sensors showing maximum amplitude of the M50 component of the event-related field. The M50 component is known to originate from S1 after tactile stimulation (Iguchi, Hoshi, Tanosaki, Taira, & Hashimoto, 2005). To identify the M50, we first averaged the time domain data for each gradiometer and each participant separately. Next, gradiometer pairs were combined by adding the signal of all trials to the two orthogonal sensors using Pythagoras' rule. The evoked responses were then averaged across participants. We identified the M50 component by focusing on the time window 0.025–0.120 sec after stimulation. Finally, we determined five sensor pairs showing maximum amplitude of the M50 (MEG1122+1123, MEG1132+1133, MEG1312+1313, MEG1342+1343, MEG1332+1333).

Regression Analyses

For each participant, we sorted the trials with intermediate SOA from low to high power, either for the gamma band or the alpha band. Then, we divided the trials in five bins with equal number of trials in each bin. There were 30.0 ± 0.1 trials per bin. Note that the sum of trials in all bins is not 200 due to trials being removed in the preprocessing steps.

To determine a potential relation between oscillatory power and perception, we determined for each bin the mean responses per participant by averaging the number of “1” and “2” responses.

For each bin, we normalized mean responses according to the following procedure (Baumgarten et al., 2016; Lange et al., 2012; Jones et al., 2010; Linkenkaer-Hansen et al., 2004): We calculated the mean response for each participant for (a) each single bin and (b) across all bins. Then, for each single bin, we subtracted the mean response across all bins from the mean response from a single bin. The obtained result was then divided by the mean response across all bins.

Finally, we calculated for each bin mean responses (and SEM) across participants.

To reproduce the results of Baumgarten et al. (2016), we performed linear regression analysis between alpha-band power and perceptual responses. To determine a potential relation between prestimulus alpha-band power and poststimulus gamma-band power, we performed regression analyses (Baumgarten et al., 2016; Lange et al., 2012; Linkenkaer-Hansen et al., 2004). Because we a priori expected a linear relationship, we first performed a linear regression. In addition, we performed a post hoc quadratic regression analysis.

To determine a potential relation between alpha-band and gamma-band power, we determined for each alpha-band power bin the average gamma-band power per participant. Next, we normalized for each participant the mean gamma-band power relative to the mean gamma-band power across all bins. Finally, we calculated for each alpha-band power bin mean gamma-band power (and SEM) across participants.

To exclude the possibility that a correlation between alpha-band power and gamma-band power was induced by covarying noise levels in both frequency bands across trials, we performed additional control analyses. To this end, we repeated the abovementioned analysis, but now with gamma-band power averaged across a different time window (but with identical length), for which we did not expect modulations of gamma-band power but just noise fluctuations (−500 to −300 msec).

Second, we computed signal-to-noise ratios (SNRs) by dividing for each participant and trial poststimulus gamma-band power (i.e., between 0 and 200 msec) and prestimulus gamma-band power (i.e., “noise” between −500 and −300 msec). Then, we repeated the abovementioned analysis for the SNRs.

All regression analyses were carried out using the Matlab built-in function regstats.

Statistical Analysis

We statistically compared perception across alpha- and gamma-band power bins, respectively. Likewise, we statistically compared gamma-band power across alpha-band power bins. First, we applied a Kolmogorov–Smirnov test to test for normality of the data for each bin. Kolmogorov–Smirnov tests showed that data in all bins significantly differed from a normal distribution (all ps < .05). To confirm and strengthen the significant linear or quadratic regression, we additionally performed planned post hoc Wilcoxon signed-ranked tests on the most extreme values, respectively. That is, for the significant linear regression between alpha-band power and perception, we compared Bins 1 and 5. For the significant quadratic regression between alpha-band power and gamma-band power, gamma-band power should be lower in alpha-band power Bin 3 relative to Bins 1 and 5. To this end, we applied one-sided Wilcoxon signed-ranked tests to compare Bin 3 versus Bin 1 and Bin 3 versus Bin 5.

RESULTS

To investigate the relationship between prestimulus alpha-band power, poststimulus gamma-band power, and perception, we measured MEG while participants performed a tactile temporal discrimination task.

Behavioral Data

Participants received one or two stimuli with varying SOAs and had to report the number of perceived stimuli. When only one stimulus was presented, participants reported one stimulus in 94.3 ± 0.4% of all trials. When two stimuli were presented with an SOA of 100 msec, participants reported two stimuli in 97.0 ± 0.3% of all trials. In addition, we presented stimuli with a predetermined individual SOA for which participants were supposed to perceive half of the trials as one stimulus and the other half as two stimuli (intermediate SOA, mean = 24.6 msec, SD = 6.2 msec). As intended, participants perceived trials with this intermediate SOA as two stimuli in 59.9 ± 0.9% of the trials. Finally, stimuli with an intermediate SOA+10 msec were perceived as two stimuli in 82.1 ± 1.3% and stimuli with an intermediate SOA-10 msec were perceived as two stimuli in 27.2 ± 1.5%.

Individual Gamma Ranges with Highest Power

We analyzed for each participant's gamma ranges with highest power within 40–150 Hz. Twelve of the 16 participants showed narrow-banded gamma-band activity within the range of 40–150 Hz (Figure 1A). Four participants showed two different gamma ranges with highest power. Three participants had to be excluded because their gamma-band activity never reached the threshold of t = 1. One participant had to be excluded because of a broadband response that extended into lower frequencies. Thus, for this participant, we could not distinguish a clear narrow-banded range of gamma-band activity.

Relation of Prestimulus Alpha and Poststimulus Gamma-band Power to Perception

We divided all trials with the intermediate SOA in five bins with respect to prestimulus alpha-band or poststimulus gamma-band power, respectively, and computed mean perception rates per bin. We found a significant negative correlation between prestimulus alpha-band power bins and perception, r(3) = 0.92, p = .03 (Figure 2A).

Figure 2. 

Regression analyses of oscillatory power and normalized temporal perceptual discrimination rate for (A) binned prestimulus alpha-band power (8–12 Hz, Bin 1 vs. Bin 5, p = .03) and (B) binned poststimulus gamma range with highest power. Insets show results of linear regression analyses (black lines). Higher number bins indicate higher spectral power. Error bars represent SEM.

Figure 2. 

Regression analyses of oscillatory power and normalized temporal perceptual discrimination rate for (A) binned prestimulus alpha-band power (8–12 Hz, Bin 1 vs. Bin 5, p = .03) and (B) binned poststimulus gamma range with highest power. Insets show results of linear regression analyses (black lines). Higher number bins indicate higher spectral power. Error bars represent SEM.

That is, with lower prestimulus alpha-band power, participants more likely reported to perceive two stimuli. Wilcoxon sign-ranked tests showed a significant difference in perception between alpha-band power Bin 1 and Bin 5 (z = 2.20, p = .03).

By contrast, we found no significant correlation between poststimulus gamma-band power and perception for both linear, r(3) = 0.04, p = .95 (Figure 2B), and quadratic, r(2) = 0.44, p = .80, regression analyses.

Relation of Prestimulus Alpha and Poststimulus Gamma-band Power

We divided all trials with the intermediate SOA in five bins with respect to prestimulus alpha-band power and computed mean gamma-band power per bin. Regression analysis did not demonstrate a significant linear relationship between prestimulus alpha-band power and poststimulus gamma-band power, r(2) = 0.22, p = .72. However, regression analysis demonstrated a significant quadratic relationship between prestimulus alpha-band power and poststimulus gamma-band power, r(2) = 0.98, p = .04 (Figure 3).

Figure 3. 

Regression analysis of binned prestimulus alpha-band power (8–12 Hz) and poststimulus gamma range with highest power. Inset shows result of quadratic regression analysis (black line). Higher number bins indicate higher spectral power. Error bars represent SEM. Bin 3 vs. Bin 1, p = .02; Bin 3 vs. Bin 5, p = .03.

Figure 3. 

Regression analysis of binned prestimulus alpha-band power (8–12 Hz) and poststimulus gamma range with highest power. Inset shows result of quadratic regression analysis (black line). Higher number bins indicate higher spectral power. Error bars represent SEM. Bin 3 vs. Bin 1, p = .02; Bin 3 vs. Bin 5, p = .03.

That is, trials with high and low prestimulus alpha-band power showed the highest poststimulus gamma-band power. Trials with intermediate prestimulus alpha-band power showed the lowest poststimulus gamma-band power.

Wilcoxon signed-rank tests revealed a significant difference in gamma-band power between alpha-band power Bins 1 and 3 (z = −2.00, p = .02), that is, bins with low prestimulus alpha-band power showed significantly higher poststimulus gamma-band power than trials with intermediate prestimulus alpha-band power. Wilcoxon sign-ranked tests also revealed a significant difference of poststimulus gamma-band power between alpha-band power Bins 3 and 5 (z = −1.84, p = .03), that is, bins with high prestimulus alpha-band power showed significantly higher poststimulus gamma power than trials with intermediate prestimulus alpha-band power. Gamma-band power in the intermediate alpha-band power bin is therefore significantly lower than in the bin with highest or lowest alpha-band power, respectively.

Control analyses revealed that this result could not be explained by common noise fluctuations in the alpha and gamma bands (Figure A1).

Figure 4 combines and summarizes the results above; with low prestimulus alpha and high poststimulus gamma-band power, participants more often perceived two stimuli. By contrast, with high poststimulus gamma-band power but with high prestimulus, alpha-band power participants more often perceived one stimulus. Finally, with intermediate alpha-band power and low poststimulus gamma-band power, participants had no clear preference for either perception (Figure 4).

Figure 4. 

Combination and summary of results. Low prestimulus alpha-band power (8–12 Hz) and high poststimulus gamma-band power lead to increased perception of two stimuli. High prestimulus alpha-band power and high poststimulus gamma-band power lead to increased perception of one stimuli. Intermediate alpha-band power and low gamma-band power lead to no clear preference for either perception.

Figure 4. 

Combination and summary of results. Low prestimulus alpha-band power (8–12 Hz) and high poststimulus gamma-band power lead to increased perception of two stimuli. High prestimulus alpha-band power and high poststimulus gamma-band power lead to increased perception of one stimuli. Intermediate alpha-band power and low gamma-band power lead to no clear preference for either perception.

DISCUSSION

We analyzed data from a previous temporal tactile discrimination task in which participants received one or two tactile stimuli with varying SOAs (Baumgarten et al., 2016). We analyzed neuronal activity recorded with MEG with respect to the relation of prestimulus alpha-band power, poststimulus gamma-band power, and tactile perception. We found a significant linear relationship between prestimulus alpha-band power and tactile perception. However, we did not find a significant correlation between poststimulus gamma-band power and tactile perception (Figure 2). Finally, we found a significant U-shaped relation between prestimulus alpha-band power and poststimulus gamma-band power (Figure 3). That is, for both lowest and highest prestimulus alpha-band power, we found the highest poststimulus gamma-band power. For intermediate prestimulus alpha-band power, we found the lowest poststimulus gamma-band power.

As in our original study (with 16 participants; Baumgarten et al., 2016), we also found a significant correlation between prestimulus alpha-band power and perception for the 12 participants in our present study. Our results are also in line with other studies reporting a linear relationship between prestimulus alpha-band power in somatosensory areas and tactile perception (Lange et al., 2012; Jones et al., 2010).

Prestimulus alpha-band power and poststimulus gamma-band power were analyzed in predefined sensors of interest. Prestimulus alpha-band power was analyzed in sensors showing a significant effect of prestimulus alpha power on perception in our previous study (Baumgarten et al., 2016). Poststimulus gamma-band power was analyzed in sensors defined by the M50 component of evoked fields. Because we performed our analyses on sensor level, we can only indirectly infer the underlying cortical sources. In our previous study, we found that the alpha effect on perception originates from somatosensory and parietal cortical regions (Baumgarten et al., 2016). In addition, the M50 component is known to originate from primary somatosensory cortex (S1; Iguchi et al., 2005). Because the poststimulus gamma response in our task strongly overlapped with the sensors defined by the M50 component (Figure 1B), it seems likely that the effect of poststimulus gamma-band activity has the same origin as the M50 event-related field component, namely, S1. This interpretation is in line with previous studies showing that poststimulus gamma-band activity in response to tactile stimulation is typically found in (primary) somatosensory areas or in sensors putatively overlying somatosensory areas (Cheng et al., 2016; Siegle et al., 2014; Lange, Oostenveld, & Fries, 2011; Gross, Schnitzler, Timmermann, & Ploner, 2007; Bauer et al., 2006). In summary, this suggests that the cortical sources of prestimulus alpha-band power and poststimulus gamma-band power might overlap but also demonstrate differences.

We focused our analysis of poststimulus gamma-band power on the time period of 0–200 msec. This time window temporally coincides with evoked activity. Such evoked activity could induce broadband activity in the frequency domain that might be misinterpreted as gamma-band activity. However, except for one participant, our analysis of the individual gamma-band ranges revealed narrow-band poststimulus gamma-band power increases that did not extend into lower frequencies (Figure 1A). We are thus confident that our gamma-band activity is not due to broadband evoked responses.

Three participants did not show a reliable range of gamma-band activity and were thus excluded from the analyses. We can only speculate about the reason for the missing gamma-band activity. One reason might be a SNR of gamma-band activity too low to be detected. Moreover, these participants showed a decrease of gamma-band power in almost all frequencies. Such a decrease is highly unusual as it indicates increased prestimulus gamma-band power relative to the poststimulus period in almost all frequencies. Because of the unusual gamma-band activity and missing gamma range with highest power (according to our criteria, see above), we thus decided to exclude these participants from further analyses.

We have analyzed gamma-band activity in the range of 40–150 Hz. Many studies have used an upper limit lower than 150 Hz for gamma oscillations oscillations (Fries, Nikolić, & Singer, 2007; Bauer et al., 2006; Hoogenboom, Schoffelen, Oostenveld, Parkes, & Fries, 2006). However, several studies have shown that gamma-band activity can extend up to 150 Hz (Lange et al., 2011; Ray, Niebur, Hsiao, Sinai, & Crone, 2008; Tallon-Baudry, Bertrand, Hénaff, Isnard, & Fischer, 2005). Therefore, we included gamma-band activity up to 150 Hz to not miss potentially important effects in the higher frequencies of the gamma band.

There has been an ongoing discussion about the nature of gamma-band oscillations. Several studies report increases of gamma-band power in narrow frequency bands in response to sensory stimulation (Krebber, Harwood, Spitzer, Keil, & Senkowski, 2015; Fries et al., 2007; Gross et al., 2007; Hoogenboom et al., 2006), arguing that gamma-band power reflects oscillatory activity. Other studies reported increases of gamma-band power in broadbands, spanning almost the entire gamma band (40 up to 200 Hz; e.g., Hermes, Miller, Wandell, & Winawer, 2015; Crone, Korzeniewska, & Franaszczuk, 2011). These studies often argue that the broadband response is unlikely of oscillatory nature but rather reflects asynchronous neuronal firing. In line with previous MEG/EEG studies, we found in our study poststimulus gamma-band responses in comparably narrow frequency bands. It seems interesting that narrow band gamma responses are often found in MEG and EEG studies, whereas broadband gamma responses are often reported in ECoG studies (e.g., Hermes et al., 2015; Lachaux et al., 2005). The nature of gamma-band power is thus far from conclusive, and thus, it is interesting and important to further elucidate the nature of gamma-band activity.

Previous studies reported increased somatosensory poststimulus gamma-band power in relation to improved tactile or nociceptive somatosensory perception (Siegle et al., 2014; Gross et al., 2007; Meador et al., 2002). Therefore, we hypothesized that poststimulus gamma-band power might correlate with perception in our tactile discrimination task. Contrary to our hypothesis, however, we did not find a significant correlation between poststimulus gamma-band power and perception. The reason for the apparent discrepancy between our study and previous studies might be found in the stimuli and tasks. Stimulus detection tasks can be near-threshold or suprathreshold. In near-threshold tasks, participants typically report whether or not they perceive a stimulus near perceptual threshold (e.g., Siegle et al., 2014; Weisz et al., 2014; van Dijk et al., 2008; Linkenkaer-Hansen et al., 2004). In suprathreshold tasks, stimuli are always above perceptual threshold, and thus, participants always perceive a stimulus but typically have to discriminate between different stimuli or perceptual states (e.g., Baumgarten et al., 2016; Peng, Hautus, Oey, & Silcock, 2016; Sato, Nagai, Kuriki, & Nakauchi, 2016; Lange et al., 2012).

Notably, the studies reporting a positive relation between poststimulus gamma-band power and perception used near-threshold stimuli and tasks. For example, detection of tactile near-threshold stimuli improved when participants exhibited higher poststimulus gamma-band power in contralateral S1 (Meador et al., 2002). Also, perceived pain around the pain threshold was accompanied by higher gamma-band power in S1 compared with unperceived pain stimuli (Gross et al., 2007). Entraining peristimulus neocortical gamma-band power optogenetically led to increased tactile stimulus detection in mice in a near-threshold detection task (Siegle et al., 2014). By contrast, we used a suprathreshold discrimination task. That is, participants always perceived a stimulus but their perception varied on a trial-by-trial basis between perceiving one or two stimuli. It has been suggested that neuronal oscillations in the gamma band are a fundamental process of neuronal communication and stimulus processing (e.g., Fries, 2005, 2015). Gamma oscillations are believed to be instrumental for efficient neuronal processing. That is, neuronal synchronization in the gamma band leads to efficient transmission of the sensory signal in the neuronal network and hence to an efficient stimulus processing (e.g., Womelsdorf & Fries, 2007). By contrast, lower gamma-band activity would then indicate that the sensory signal is transmitted less efficiently across the neuronal network and hence the signal is less efficiently processed, leading potentially to a less clear and potentially even ambiguous perception. In line with this hypothesis, low gamma-band power in a near-threshold detection task might indicate that the stimulus is insufficiently processed and thus not perceived. By contrast, high gamma-band power indicates efficient stimulus processing, leading to successful detection of the near-threshold stimulus (Siegle et al., 2014; Gross et al., 2007). In suprathreshold tasks, a stimulus is always strong enough to be sufficiently processed to result in successful perception. Therefore, a suprathreshold task should display high gamma-band power for all stimuli.

In our study, we used stimuli with identical physical characteristics (two suprathreshold stimuli with intermediate SOA), which differed only in participants' subjective perception. Gamma-band power was present in all trials, indicating efficient stimulus processing. However, the lack of a significant difference in gamma-band power between perceiving one or two stimuli suggests that the stimulus processing in S1 is largely independent of subjective perception in suprathreshold tasks. Subjective perception might be processed in other, higher cortical areas. For example, studies using working memory tasks in humans and monkeys found that vibrotactile stimulation induced gamma-band power in somatosensory areas. Somatosensory gamma-band power, however, did not differ between correctly and incorrectly perceived trials. Such differences between subjective perception and gamma-band power were found in higher areas (Haegens, Nácher, Hernández, et al., 2011; Haegens et al., 2010).

An alternative explanation for the lack of a significant correlation between poststimulus gamma-band power and perception might be that a potential correlation between gamma-band power and subjective perception might be too small to be detected with our paradigm or analysis approach. In addition, differences in gamma-band power might occur at different frequencies than analyzed in our study. However, we focused our analysis on individual frequency bands showing gamma-band power in response to stimulation, whereas other frequency bands showed only negligible gamma-band power, at all.

In contrast to our study in the somatosensory domain, studies in the visual domain reported that poststimulus gamma-band power correlated with subjective perception in suprathreshold tasks. These differences in gamma-band power, however, were typically found in higher visual areas, other than primary visual cortex. For example, if participants receive one visual stimulus accompanied by two tactile stimuli, they frequently perceive a second illusory visual stimulus (Shams, Kamitani, & Shimojo, 2000).

Studies have shown that, despite identical physical stimulation, poststimulus gamma-band power in parieto-occipital cortex correlated with participants' subjective perception of the illusion (Balz et al., 2016; Lange et al., 2011; Bhattacharya, Shams, & Shimojo, 2002). Moreover, poststimulus gamma-band power in somatosensory cortices was larger for congruent compared with incongruent visuotactile stimuli and correlated with shorter RTs (Krebber et al., 2015). Future studies might thus further investigate how gamma-band power correlates with tactile perception in suprathreshold tasks by studying other cortical areas or using methodological approaches that allow a finer spatial resolution, such as intracranial EEG or local field potential recording.

The main focus of our study was to study a potential relationship between prestimulus alpha and poststimulus gamma-band power. It has been shown that attention correlates negatively with prestimulus alpha-band power and positively with poststimulus gamma-band power in somatosensory areas (Haegens, Luther, & Jensen, 2012; Haegens, Nácher, Luna, Romo, & Jensen, 2011; Bauer et al., 2006). In addition, higher behavioral performance is associated with lower prestimulus alpha-band power and higher poststimulus gamma-band power (e.g., Baumgarten et al., 2016; Siegle et al., 2014). We thus hypothesized that prestimulus alpha and poststimulus gamma-band power negatively correlate on a trial-by-trial basis, a question that to our knowledge has never been directly investigated. In contrast to our hypothesis, we did not find a significant linear relationship. Rather, we found that prestimulus alpha and poststimulus gamma-band power show a quadratic relationship. That is, low but also high prestimulus alpha-band power was associated with high poststimulus gamma-band power, whereas intermediate levels of prestimulus alpha-band power were associated with low levels of poststimulus gamma-band power. In addition, in trials with low prestimulus alpha/high poststimulus gamma-band power, participants more often perceived two stimuli, whereas in trials with high prestimulus alpha/high poststimulus gamma-band power, participants perceived more often one stimulus (Figure 4). Furthermore, in trials with intermediate prestimulus alpha/low poststimulus gamma-band power, participants showed no preference for either perception.

Although this quadratic relation was shown to be significant, the overall effect sizes seem rather small. We can only speculate about the size of the effects. It might be that only a small fraction of neurons that elicit gamma-band activity are involved in the perception process and are modulated by prestimulus alpha-band power. This would lead to a comparably low SNR and thus small effect sizes. Another potential reason might be found in the overall lower SNR for higher frequencies. Such a low SNR might reduce potential effects. The effect sizes in our study are, however, comparable in size to effect sizes of gamma-band effects in other MEG studies (Yuan, Li, Liu, Yuan, & Huang, 2016; Krebber et al., 2015; Haegens et al., 2010).

We propose that low prestimulus alpha-band power reflects states of high excitability (Iemi et al., 2017; Lange et al., 2013; Thut et al., 2006). Therefore, stimuli will be efficiently processed during states of low prestimulus alpha-band power, resulting in the perception of two stimuli (Baumgarten et al., 2016).

The lower prestimulus alpha-band power, the higher was participants' confidence in their decision. In other words, stronger or more efficient processing of “two” stimuli is accompanied by lower alpha-band power (Baumgarten et al., 2016).

Such efficient stimulus processing should be reflected in high poststimulus gamma-band power (Fries, 2005, 2009).

Hence, we propose that low prestimulus alpha-band power will lead to high poststimulus gamma-band power, resulting in the perception of two stimuli (Figure 4, upper curve). On the other hand, high prestimulus alpha-band power reflects lower excitability or pulsed inhibition (Jensen & Mazaheri, 2010; Mathewson, Gratton, Fabiani, Beck, & Ro, 2009), leading to the perception of only one stimulus (Baumgarten et al., 2016).

The higher prestimulus alpha-band power, the higher was participants' confidence in their decision of “one” stimulus. In other words, stronger or more efficient processing of “one” stimuli was accompanied by higher alpha-band power (Baumgarten et al., 2016). Again, such efficient stimulus processing (despite leading to erroneous perception) should be reflected in high poststimulus gamma-band power (Fries, 2005, 2009). Thus, we propose that high prestimulus alpha-band power should also lead to high poststimulus gamma-band power. This way, however, high gamma-band power will result in the perception of one stimulus (Figure 4, lower curve). Finally, intermediate level of prestimulus alpha-band power will not bias perception in either direction, leading to lower or inefficient forwarding of the stimulus, which will be reflected in lower levels of gamma-band power.

This proposed model offers an alternative explanation why we did not find a significant correlation between gamma-band power and perception (Figure 2B). If prestimulus alpha-band power determines whether high poststimulus gamma-band power reflects the perception of one or two stimuli, then averaging across all prestimulus alpha states (as done in Figure 2B) will also average across both perceptions. Thus, ignoring the prestimulus alpha state and simply looking at poststimulus gamma states might give the wrong impression of no correlation between poststimulus gamma-band power and perception.

In conclusion, we found that prestimulus alpha-band and poststimulus gamma-band power show a quadratic relationship with both low and high prestimulus alpha power, leading to high poststimulus gamma-band power. Notably, the two states of high poststimulus gamma-band power are related to different states of perception. We propose a model in which prestimulus alpha-band power determines the computational and perceptual fate of a stimulus. If prestimulus alpha-band power is low, stimuli are efficiently processed, leading to more veridical perception in suprathreshold temporal discrimination tasks or near-threshold detection tasks. In such cases, poststimulus gamma-band power will be high, indicating efficient stimulus processing. If prestimulus alpha-band power is high, stimuli are inefficiently processed, leading to more incorrect perceptions in suprathreshold temporal discrimination tasks and no perception in near-threshold detection tasks. In suprathreshold temporal discrimination tasks, stimuli will still be processed, leading to high gamma-band power. In near-threshold detection task, nonperceived stimuli will not be processed, leading to no poststimulus gamma-band power.

APPENDIX

Figure A1. 

Regression analysis of binned prestimulus alpha-band power (8–12 Hz) and (A) prestimulus gamma-band noise or (B) SNR of gamma-band power. Inset shows results of quadratic regression analyses (black line). Higher number bins indicate higher spectral power. Error bars represent SEM. For (A): Bin 1 versus Bin 3, p = .95; Bin 3 versus Bin 5, p = .48. For (B): Bin 3 versus Bin 1, p = .026; Bin 3 versus Bin 5, p = .002.

Figure A1. 

Regression analysis of binned prestimulus alpha-band power (8–12 Hz) and (A) prestimulus gamma-band noise or (B) SNR of gamma-band power. Inset shows results of quadratic regression analyses (black line). Higher number bins indicate higher spectral power. Error bars represent SEM. For (A): Bin 1 versus Bin 3, p = .95; Bin 3 versus Bin 5, p = .48. For (B): Bin 3 versus Bin 1, p = .026; Bin 3 versus Bin 5, p = .002.

Acknowledgments

J. L. was supported by the German Research Foundation (LA 2400/4-1).

Reprint requests should be sent to Marc André Wittenberg, Institute of Clinical Neuroscience and Medical Psychology, Medical Faculty, Heinrich-Heine-University Düsseldorf, Universitätsstrasse 1, Düsseldorf, 40225, Germany, or via e-mail: marc.wittenberg@hhu.de.

REFERENCES

REFERENCES
Balz
,
J.
,
Keil
,
J.
,
Roa Romero
,
Y.
,
Mekle
,
R.
,
Schubert
,
F.
,
Aydin
,
S.
, et al
(
2016
).
GABA concentration in superior temporal sulcus predicts gamma power and perception in the sound-induced flash illusion
.
Neuroimage
,
125
,
724
730
.
Bauer
,
M.
,
Oostenveld
,
R.
,
Peeters
,
M.
, &
Fries
,
P.
(
2006
).
Tactile spatial attention enhances gamma-band activity in somatosensory cortex and reduces low-frequency activity in parieto-occipital areas
.
Journal of Neuroscience
,
26
,
490
501
.
Baumgarten
,
T. J.
,
Schnitzler
,
A.
, &
Lange
,
J.
(
2016
).
Prestimulus alpha power influences tactile temporal perceptual discrimination and confidence in decisions
.
Cerebral Cortex
,
26
,
891
903
.
Baumgarten
,
T. J.
,
Schnitzler
,
A.
, &
Lange
,
J.
(
2017
).
Beyond the peak—Tactile temporal discrimination does not correlate with individual peak frequencies in somatosensory cortex
.
Frontiers in Psychology
,
8
,
421
.
Bhattacharya
,
J.
,
Shams
,
L.
, &
Shimojo
,
S.
(
2002
).
Sound-induced illusory flash perception: Role of gamma band responses
.
NeuroReport
,
13
,
1727
1730
.
Buzsáki
,
G.
, &
Draguhn
,
A.
(
2004
).
Neuronal oscillations in cortical networks
.
Science
,
304
,
1926
1929
.
Buzsáki
,
G.
, &
Watson
,
B. O.
(
2012
).
Brain rhythms and neural syntax: Implications for efficient coding of cognitive content and neuropsychiatric disease
.
Dialogues in Clinical Neuroscience
,
14
,
345
367
.
Cheng
,
C.-H.
,
Chan
,
P.-Y. S.
,
Niddam
,
D. M.
,
Tsai
,
S.-Y.
,
Hsu
,
S.-C.
, &
Liu
,
C.-Y.
(
2016
).
Sensory gating, inhibition control and gamma oscillations in the human somatosensory cortex
.
Scientific Reports
,
6
,
20437
.
Cousijn
,
H.
,
Haegens
,
S.
,
Wallis
,
G.
,
Near
,
J.
,
Stokes
,
M. G.
,
Harrison
,
P. J.
, et al
(
2014
).
Resting GABA and glutamate concentrations do not predict visual gamma frequency or amplitude
.
Proceedings of the National Academy of Sciences, U.S.A.
,
111
,
9301
9306
.
Crone
,
N. E.
,
Korzeniewska
,
A.
, &
Franaszczuk
,
P. J.
(
2011
).
Cortical γ responses: Searching high and low
.
International Journal of Psychophysiology
,
79
,
9
15
.
Foxe
,
J. J.
,
Simpson
,
G. V.
, &
Ahlfors
,
S. P.
(
1998
).
Parieto-occipital approximately 10 Hz activity reflects anticipatory state of visual attention mechanisms
.
NeuroReport
,
9
,
3929
3933
.
Fries
,
P.
(
2005
).
A mechanism for cognitive dynamics: Neuronal communication through neuronal coherence
.
Trends in Cognitive Sciences
,
9
,
474
480
.
Fries
,
P.
(
2009
).
Neuronal gamma-band synchronization as a fundamental process in cortical computation
.
Annual Review of Neuroscience
,
32
,
209
224
.
Fries
,
P.
(
2015
).
Rhythms for cognition: Communication through coherence
.
Neuron
,
88
,
220
235
.
Fries
,
P.
,
Nikolić
,
D.
, &
Singer
,
W.
(
2007
).
The gamma cycle
.
Trends in Neurosciences
,
30
,
309
316
.
Fries
,
P.
,
Womelsdorf
,
T.
,
Oostenveld
,
R.
, &
Desimone
,
R.
(
2008
).
The effects of visual stimulation and selective visual attention on rhythmic neuronal synchronization in macaque area V4
.
Journal of Neuroscience
,
28
,
4823
4835
.
Gross
,
J.
,
Schnitzler
,
A.
,
Timmermann
,
L.
, &
Ploner
,
M.
(
2007
).
Gamma oscillations in human primary somatosensory cortex reflect pain perception
.
PLoS Biology
,
5
,
e133
.
Haegens
,
S.
,
Luther
,
L.
, &
Jensen
,
O.
(
2012
).
Somatosensory anticipatory alpha activity increases to suppress distracting input
.
Journal of Cognitive Neuroscience
,
24
,
677
685
.
Haegens
,
S.
,
Nácher
,
V.
,
Hernández
,
A.
,
Luna
,
R.
,
Jensen
,
O.
, &
Romo
,
R.
(
2011
).
Beta oscillations in the monkey sensorimotor network reflect somatosensory decision making
.
Proceedings of the National Academy of Sciences, U.S.A.
,
108
,
10708
10713
.
Haegens
,
S.
,
Nácher
,
V.
,
Luna
,
R.
,
Romo
,
R.
, &
Jensen
,
O.
(
2011
).
α-Oscillations in the monkey sensorimotor network influence discrimination performance by rhythmical inhibition of neuronal spiking
.
Proceedings of the National Academy of Sciences, U.S.A.
,
108
,
19377
19382
.
Haegens
,
S.
,
Osipova
,
D.
,
Oostenveld
,
R.
, &
Jensen
,
O.
(
2010
).
Somatosensory working memory performance in humans depends on both engagement and disengagement of regions in a distributed network
.
Human Brain Mapping
,
31
,
26
35
.
Händel
,
B. F.
,
Haarmeier
,
T.
, &
Jensen
,
O.
(
2011
).
Alpha oscillations correlate with the successful inhibition of unattended stimuli
.
Journal of Cognitive Neuroscience
,
23
,
2494
2502
.
Hanslmayr
,
S.
,
Aslan
,
A.
,
Staudigl
,
T.
,
Klimesch
,
W.
,
Herrmann
,
C. S.
, &
Bäuml
,
K.-H.
(
2007
).
Prestimulus oscillations predict visual perception performance between and within subjects
.
Neuroimage
,
37
,
1465
1473
.
Hermes
,
D.
,
Miller
,
K. J.
,
Wandell
,
B. A.
, &
Winawer
,
J.
(
2015
).
Stimulus dependence of gamma oscillations in human visual cortex
.
Cerebral Cortex
,
25
,
2951
2959
.
Hoogenboom
,
N.
,
Schoffelen
,
J.-M.
,
Oostenveld
,
R.
, &
Fries
,
P.
(
2010
).
Visually induced gamma-band activity predicts speed of change detection in humans
.
Neuroimage
,
51
,
1162
1167
.
Hoogenboom
,
N.
,
Schoffelen
,
J.-M.
,
Oostenveld
,
R.
,
Parkes
,
L. M.
, &
Fries
,
P.
(
2006
).
Localizing human visual gamma-band activity in frequency, time and space
.
Neuroimage
,
29
,
764
773
.
Iemi
,
L.
,
Chaumon
,
M.
,
Crouzet
,
S. M.
, &
Busch
,
N. A.
(
2017
).
Spontaneous neural oscillations bias perception by modulating baseline excitability
.
Journal of Neuroscience
,
37
,
807
819
.
Iguchi
,
Y.
,
Hoshi
,
Y.
,
Tanosaki
,
M.
,
Taira
,
M.
, &
Hashimoto
,
I.
(
2005
).
Attention induces reciprocal activity in the human somatosensory cortex enhancing relevant- and suppressing irrelevant inputs from fingers
.
Clinical Neurophysiology
,
116
,
1077
1087
.
Jensen
,
O.
, &
Mazaheri
,
A.
(
2010
).
Shaping functional architecture by oscillatory alpha activity: Gating by inhibition
.
Frontiers in Human Neuroscience
,
4
,
186
.
Jones
,
S. R.
,
Kerr
,
C. E.
,
Wan
,
Q.
,
Pritchett
,
D. L.
,
Hämäläinen
,
M.
, &
Moore
,
C. I.
(
2010
).
Cued spatial attention drives functionally relevant modulation of the mu rhythm in primary somatosensory cortex
.
Journal of Neuroscience
,
30
,
13760
13765
.
Keil
,
J.
,
Müller
,
N.
,
Ihssen
,
N.
, &
Weisz
,
N.
(
2012
).
On the variability of the McGurk effect: Audiovisual integration depends on prestimulus brain states
.
Cerebral Cortex
,
22
,
221
231
.
Klimesch
,
W.
,
Sauseng
,
P.
, &
Hanslmayr
,
S.
(
2007
).
EEG alpha oscillations: The inhibition-timing hypothesis
.
Brain Research Reviews
,
53
,
63
88
.
Krebber
,
M.
,
Harwood
,
J.
,
Spitzer
,
B.
,
Keil
,
J.
, &
Senkowski
,
D.
(
2015
).
Visuotactile motion congruence enhances gamma-band activity in visual and somatosensory cortices
.
Neuroimage
,
117
,
160
169
.
Lachaux
,
J.-P.
,
George
,
N.
,
Tallon-Baudry
,
C.
,
Martinerie
,
J.
,
Hugueville
,
L.
,
Minotti
,
L.
, et al
(
2005
).
The many faces of the gamma band response to complex visual stimuli
.
Neuroimage
,
25
,
491
501
.
Lange
,
J.
,
Halacz
,
J.
,
van Dijk
,
H.
,
Kahlbrock
,
N.
, &
Schnitzler
,
A.
(
2012
).
Fluctuations of prestimulus oscillatory power predict subjective perception of tactile simultaneity
.
Cerebral Cortex
,
22
,
2564
2574
.
Lange
,
J.
,
Keil
,
J.
,
Schnitzler
,
A.
,
van Dijk
,
H.
, &
Weisz
,
N.
(
2014
).
The role of alpha oscillations for illusory perception
.
Behavioural Brain Research
,
271
,
294
301
.
Lange
,
J.
,
Oostenveld
,
R.
, &
Fries
,
P.
(
2011
).
Perception of the touch-induced visual double-flash illusion correlates with changes of rhythmic neuronal activity in human visual and somatosensory areas
.
Neuroimage
,
54
,
1395
1405
.
Lange
,
J.
,
Oostenveld
,
R.
, &
Fries
,
P.
(
2013
).
Reduced occipital alpha power indexes enhanced excitability rather than improved visual perception
.
Journal of Neuroscience
,
33
,
3212
3220
.
Linkenkaer-Hansen
,
K.
,
Nikulin
,
V. V.
,
Palva
,
S.
,
Ilmoniemi
,
R. J.
, &
Palva
,
J. M.
(
2004
).
Prestimulus oscillations enhance psychophysical performance in humans
.
Journal of Neuroscience
,
24
,
10186
10190
.
Mathewson
,
K. E.
,
Gratton
,
G.
,
Fabiani
,
M.
,
Beck
,
D. M.
, &
Ro
,
T.
(
2009
).
To see or not to see: Prestimulus alpha phase predicts visual awareness
.
Journal of Neuroscience
,
29
,
2725
2732
.
Meador
,
K. J.
,
Ray
,
P. G.
,
Echauz
,
J. R.
,
Loring
,
D. W.
, &
Vachtsevanos
,
G. J.
(
2002
).
Gamma coherence and conscious perception
.
Neurology
,
59
,
847
854
.
Müller
,
M. M.
,
Gruber
,
T.
, &
Keil
,
A.
(
2000
).
Modulation of induced gamma band activity in the human EEG by attention and visual information processing
.
International Journal of Psychophysiology
,
38
,
283
299
.
Oostenveld
,
R.
,
Fries
,
P.
,
Maris
,
E.
, &
Schoffelen
,
J.-M.
(
2011
).
FieldTrip: Open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data
.
Computational Intelligence and Neuroscience
,
2011
,
156869
.
Peng
,
M.
,
Hautus
,
M. J.
,
Oey
,
I.
, &
Silcock
,
P.
(
2016
).
Is there a generalized sweetness sensitivity for an individual? A psychophysical investigation of inter-individual differences in detectability and discriminability for sucrose and fructose
.
Physiology & Behavior
,
165
,
239
248
.
Ray
,
S.
,
Niebur
,
E.
,
Hsiao
,
S. S.
,
Sinai
,
A.
, &
Crone
,
N. E.
(
2008
).
High-frequency gamma activity (80–150 Hz) is increased in human cortex during selective attention
.
Clinical Neurophysiology
,
119
,
116
133
.
Sato
,
T.
,
Nagai
,
T.
,
Kuriki
,
I.
, &
Nakauchi
,
S.
(
2016
).
Dissociation of equilibrium points for color-discrimination and color-appearance mechanisms in incomplete chromatic adaptation
.
Journal of the Optical Society of America. A, Optics, Image Science, and Vision
,
33
,
A150
A163
.
Shams
,
L.
,
Kamitani
,
Y.
, &
Shimojo
,
S.
(
2000
).
Illusions. What you see is what you hear
.
Nature
,
408
,
788
.
Siegel
,
M.
,
Donner
,
T. H.
,
Oostenveld
,
R.
,
Fries
,
P.
, &
Engel
,
A. K.
(
2008
).
Neuronal synchronization along the dorsal visual pathway reflects the focus of spatial attention
.
Neuron
,
60
,
709
719
.
Siegle
,
J. H.
,
Pritchett
,
D. L.
, &
Moore
,
C. I.
(
2014
).
Gamma-range synchronization of fast-spiking interneurons can enhance detection of tactile stimuli
.
Nature Neuroscience
,
17
,
1371
1379
.
Tallon-Baudry
,
C.
,
Bertrand
,
O.
,
Hénaff
,
M.-A.
,
Isnard
,
J.
, &
Fischer
,
C.
(
2005
).
Attention modulates gamma-band oscillations differently in the human lateral occipital cortex and fusiform gyrus
.
Cerebral Cortex
,
15
,
654
662
.
Thut
,
G.
,
Nietzel
,
A.
,
Brandt
,
S. A.
, &
Pascual-Leone
,
A.
(
2006
).
Alpha-band electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts visual target detection
.
Journal of Neuroscience
,
26
,
9494
9502
.
van Dijk
,
H.
,
Schoffelen
,
J.-M.
,
Oostenveld
,
R.
, &
Jensen
,
O.
(
2008
).
Prestimulus oscillatory activity in the alpha band predicts visual discrimination ability
.
Journal of Neuroscience
,
28
,
1816
1823
.
Weisz
,
N.
,
Wühle
,
A.
,
Monittola
,
G.
,
Demarchi
,
G.
,
Frey
,
J.
,
Popov
,
T.
, et al
(
2014
).
Prestimulus oscillatory power and connectivity patterns predispose conscious somatosensory perception
.
Proceedings of the National Academy of Sciences, U.S.A.
,
111
,
E417
E425
.
Womelsdorf
,
T.
, &
Fries
,
P.
(
2007
).
The role of neuronal synchronization in selective attention
.
Current Opinion in Neurobiology
,
17
,
154
160
.
Womelsdorf
,
T.
,
Fries
,
P.
,
Mitra
,
P. P.
, &
Desimone
,
R.
(
2006
).
Gamma-band synchronization in visual cortex predicts speed of change detection
.
Nature
,
439
,
733
736
.
Worden
,
M. S.
,
Foxe
,
J. J.
,
Wang
,
N.
, &
Simpson
,
G. V.
(
2000
).
Anticipatory biasing of visuospatial attention indexed by retinotopically specific alpha-band electroencephalography increases over occipital cortex
.
Journal of Neuroscience
,
20
,
RC63
.
Yuan
,
X.
,
Li
,
H.
,
Liu
,
P.
,
Yuan
,
H.
, &
Huang
,
X.
(
2016
).
Pre-stimulus beta and gamma oscillatory power predicts perceived audiovisual simultaneity
.
International Journal of Psychophysiology
,
107
,
29
36
.