Differential impact of movement on the alpha and gamma dynamics serving visual processing

Abstract

Visual processing is widely understood to be served by a decrease in alpha activity in occipital cortices, largely concurrent with an increase in gamma activity. Although the characteristics of these oscillations are well documented in response to a range of complex visual stimuli, little is known about how these dynamics are impacted by concurrent motor responses, which is problematic as many common visual tasks involve such responses. Thus, in the current study, we used magnetoencephalography (MEG) and modified a well-established visual paradigm to explore the impact of motor responses on visual oscillatory activity. Thirty-four healthy adults viewed a moving gabor (grating) stimulus that was known to elicit robust alpha and gamma oscillations in occipital cortices. Frequency and power characteristics were assessed statistically for differences as a function of movement condition. Our results indicated that occipital alpha significantly increased in power during movement relative to no movement trials. No differences in peak frequency or power were found for gamma responses between the two movement conditions. These results provide valuable evidence of visuomotor integration and underscore the importance of careful task design and interpretation, especially in the context of complex visual processing, and suggest that even basic motor responses alter occipital visual oscillations in healthy adults.

NEW & NOTEWORTHY Processing of visual stimuli is served by occipital alpha and gamma activity. Many studies have investigated the impact of visual stimuli on motor cortical responses, but few studies have systematically investigated the impact of motor responses on visual oscillations. We found that when participants are asked to move in response to a visual stimulus, occipital alpha power was modulated whereas gamma responses were unaffected. This suggests that these responses have dissociable roles in visuomotor integration.

INTRODUCTION

Fast, rhythmically bursting networks of neurons have been shown to intrinsically oscillate at low amplitudes, and robustly increase in amplitude in response to sensory stimulation. It has been suggested that the modulation of band-constrained neural oscillatory activity plays a major role in information processing. With regards to visual information processing, it has long been established that sub-populations of neurons responding to different aspects of a complex visual presentation oscillate synchronously in the gamma band (1–5). Early evidence from microelectrode recordings in macaques suggested that induced local gamma event-related synchronization (gamma ERS) may be a mechanism for neuronal group communication by which different aspects of a complex stimulus can be “bound” together to perceive a complex object (3, 5). Induced gamma band responses have also been shown to be crucial to an array of cognitive functions including attention (6–14), working memory (7, 15, 16), and production of motor responses (17–20).

In addition to the visual gamma ERS response, alpha event-related desynchronization (alpha ERD) responses have been consistently associated with various aspects of visual processing (10, 12, 21–24). Desynchronization in the alpha band has thus come to be interpreted as the cortical disinhibition in response to incoming visual stimuli and visual cognitive processing (25–30). Recent studies in EEG and magnetoencephalography (MEG) have expanded this traditional view of the role of alpha activity to include a wealth of other high-order cognitive operations. For example, alpha band activity has since been recognized as an active functional mechanism for motor planning (31, 32), attention (10, 26, 33–36), and working memory (16, 37–42).

In recent years, a number of novel task paradigms and analysis methods have been implemented to dissociate the impact of various visual stimulus properties on oscillatory activity. It has been found that spectral power and frequency characteristics of the visual gamma ERS response are dependent on the properties of the presented stimulus, such as contrast and spatial frequency (43, 44), stimulus size (45–47), and motion parameters (47, 48). For example, Muthukumaraswamy et al. (47) found that annular stimuli produced greater gamma amplitude than gabor (grating) stimuli, moving gabor patterns elicited gamma oscillations of higher amplitude and frequency than static stimuli, and that gamma oscillatory amplitude was modulated by stimulus size. In contrast, the frequency and amplitude of alpha oscillations in the visual cortices were not found to be modulated by characteristics of visual stimuli (48). Swettenham et al. (48) thus suggest that alpha band amplitude is dependent on directed attention to a visual stimulus, and not the stimulus properties themselves, while gamma amplitude seems to be modulated by both. It is also important to note that, although alpha and gamma have been shown to be differentially modulated by properties of visual input, there is remarkable within-subject consistency on the spectral and amplitude characteristics of these oscillations. However, the between-subject variability in these responses is generally relatively high (49).

Given the sensitivity of gamma oscillatory activity to task parameters, as well as the broad variability in these visual responses between individuals, it is paramount that we understand the impact of extraneous variables to accurately analyze these stimulus effects. One unexplored parameter in visual alpha and gamma production is the implementation of movement as a response to task parameters. Although many studies have already taken the precaution of requiring movement responses at the end of visual stimuli (rather than at stimulus onset) to avoid potential motor contamination (11, 19, 47), to our knowledge the effect of movement on production of visual oscillatory activity has never been systematically studied. In contrast, many studies using go/no-go tasks and similar designs have investigated the effect of visual cues on movement-related oscillatory activity (i.e., the other way around) (50, 51). Of note, these studies generally see reduced alpha ERD (i.e., less different than baseline) in sensorimotor regions in the no-go condition relative to the go condition (51), but the absence of movement has a large effect, and this paradigm has not been studied in the context of visual oscillations. This is especially pertinent, as both gamma and alpha oscillatory activity has been shown to be important to both visual and motor processing (52, 53). Furthermore, mechanistic investigations have found that visuomotor integration tasks simultaneously activate both primary motor and visual regions and that connectivity between these regions is tightly coupled to an individual’s genetic profile (54). Thus, one would expect perturbations in the motor system to also affect specific visual oscillatory responses; however, this has yet to be systematically studied. Hence, the goal of the current study was to investigate how movement modulates the occipital alpha and gamma oscillatory dynamics serving visual processing. To this end, we used MEG and a visuomotor task to investigate the impact of movement on cortical activity during visual processing. We hypothesized that both alpha and gamma power and peak frequency would be significantly modulated by movement during the visual cue.

METHODS

Participant Selection

We studied 34 healthy participants (15 females; 2 left-handed), all of whom were recruited from the local community. The mean age was 26.34 yr, with a range of 19–37 yr. Exclusionary criteria included any medical illness affecting CNS function (e.g., HIV/AIDS), neurological or psychiatric disorder, history of head trauma, current substance abuse, and the MEG laboratory’s standard exclusion criteria (e.g., any type of ferromagnetic implanted material). After complete description of the study was given to participants, written informed consent was obtained following the guidelines of the University of Nebraska Medical Center’s Institutional Review Board, which approved the study protocol.

Experimental Paradigm and Stimuli

During MEG recording, participants were seated in a nonmagnetic chair within the magnetically shielded room (MSR) with the lights off, and each participant rested their right hand on a custom-made button pad. This response pad was connected such that each button sent a unique signal (i.e., TTL pulse/trigger code) to the MEG system acquisition computer, and thus behavioral responses were temporally synced with the MEG data. Participants were instructed to remain still and fixate on a small red crosshair presented centrally. After a randomly varied fixation period of 2.0–2.5 s, a moving gabor grating stimulus appeared on the screen. The gabor stimulus was largely similar to those used to elicit visual alpha and gamma in other studies (48, 50, 51). Briefly, the stimulus consisted of a 15-cycle, 100% luminance contrast grating that was rotated 135° counterclockwise from the horizontal, with a spatial frequency of 0.85 cycles/° and sigma of Gaussian of 2.96 degrees, subtended a visual angle of 16.93°. The implied motion of the gabor stimulus initially appeared at 4.2°/s toward the upper right corner of the screen. After 1.0 s, the stimulus began to move at either 2.4°/s (i.e., decreased speed) or 6.0°/s (i.e., increased speed), and remained at this speed for 1.0 s. Participants were asked to indicate with a button press as quickly as possible whenever one of the two speed changes occurred (i.e., either when the gabor stimulus increased or decreased in speed). The condition to which participants responded (i.e., increase/decrease in speed) was counterbalanced so that there were no systematic differences in the characteristics of the gabor stimulus that were being responded to across participants. In other words, half of the participants responded by button press (i.e., “Move” condition) only if the stimulus increased in speed and not if the stimulus decreased in speed (i.e., “No Move” condition), whereas the other half responded only if the stimulus decreased in speed (“Move” condition) and not when the increased in speed (“No Move” condition). An example trial is shown in Fig. 1. The order of trials was pseudorandomized such that no more than three trials of the same condition were presented in a row, and participants completed a total of 200 trials (100 of each speed). The experiment lasted ∼17 min.

Figure 1.

Visual task paradigm. The fixation cross was presented 2.0–2.5 s before the mobile-gabor patch is presented. After 1.0 s, the gabor grating then either sped up or slowed down for 1.0 s. Participants were instructed to respond by button press to either an increase or decrease in the speed of the grating stimulus; the condition to which they responded was counter-balanced between participants.

MEG Data Acquisition and Coregistration with Structural MRI

All recordings were conducted in a one-layer magnetically shielded room with active shielding engaged. Neuromagnetic responses were sampled continuously at 1 kHz with an acquisition bandwidth of 0.1–330 Hz using a 306-sensor Elekta MEG system (Elekta, Helsinki, Finland). MEG data from each individual were corrected for head motion and subjected to noise reduction using the signal space separation method with a temporal extension (tSSS) (55, 56). Each participant’s MEG data were then coregistered with a high-resolution structural T1-weighted template MRI before the application of source space analyses (i.e., beamforming) using BESA MRI (v. 2.0). The structural volume was aligned parallel to the anterior and posterior commissures and transformed into standardized space.

MEG Preprocessing, Time-Frequency Transformation, and Sensor-Level Statistics

Cardiac and blink artifacts were removed from the data using signal-space projection (SSP), which was accounted for during source reconstruction (57). The continuous magnetic time series was divided into epochs of 4.0-s duration (−1.0 s to 3.0 s, 0.0 s = stimulus onset), with the baseline defined as −0.7 to -0.2 s before stimulus onset. Epochs containing artifacts were rejected based on a fixed threshold method, supplemented with visual inspection. Artifact-free epochs were transformed into the time-frequency domain using a complex demodulation approach. Briefly, complex demodulation works by first transforming the signal into the frequency space, using a Fast Fourier Transform (FFT). This results in a frequency spectrum, inherently containing the same power and cross-spectrum information as the original signal. From here, this frequency spectrum is (de)modulated in a step-wise manner to adopt the center frequency of a series of complex sinusoids with increasing carrier frequencies, in a process termed heterodyning. These resulting signals are then low-pass filtered with a finite impulse response (FIR) filter at 4 Hz (corresponding to a full-width half-max in the temporal domain of 130 ms) (58) to reduce spectral leakage; the nature of this filter inherently determines the time and frequency resolution of the resulting data. For this study, the time-frequency analysis was performed with a frequency step of 2 Hz and a time-step of 25 ms from 4 to 100 Hz (59–61). The resulting spectral power estimations per sensor were averaged over trials to generate time-frequency plots of mean spectral density. These sensor-level data were normalized per time-frequency bin by subtracting baseline power (calculated as the mean power during the −0.7 s to −0.2 s prestimulus time period) from each bin, then dividing by the baseline power for that bin and multiplying by 100 to yield a percentage change from baseline {i.e., [(bin power – baseline power)/baseline power] × 100}.

The specific time-frequency windows used for imaging were determined by statistical analysis of the sensor-level spectrograms collapsed across both conditions across the entire array of gradiometers and all participants. Each data point in the spectrogram was initially evaluated using a mass univariate approach based on the general linear model (GLM). To reduce the risk of false positive results while maintaining reasonable sensitivity, a two-stage procedure was followed to control for type 1 error. In the first stage, paired-sample t tests against baseline were conducted on each data point and the output spectrogram of t values was thresholded at P < 0.05 to define time-frequency bins containing potentially significant oscillatory deviations across all participants. In the second stage, time-frequency bins that survived the threshold were clustered with temporally and/or spectrally neighboring bins that were also below the (P < 0.05) threshold and a cluster value was derived by summing all of the t values of all data points in the cluster. Nonparametric permutation testing was then used to derive a distribution of cluster values, and the significance level of the observed clusters (from stage 1) were tested directly using this distribution (62, 63). For each comparison, 1,000 permutations were computed to build a distribution of cluster values. Based on these analyses, time-frequency windows that contained a significant oscillatory event across all participants and conditions (e.g., alpha ERD, gamma ERS) were subjected to the beamforming analysis. Of note, the number of accepted trials included in analysis were not significantly different between conditions [Move: 80.94 (SD: 7.31) trials, No-Move: 81.65 (SD: 7.81) trials; P > 0.05].

MEG Imaging, Virtual Sensor Extraction, and Statistics

Cortical networks were imaged through an extension of the linearly constrained minimum variance vector beamformer (64, 65), which employs spatial filters in the frequency domain to calculate source power for the entire brain volume. The images were derived from the cross-spectral densities of all combinations of MEG gradiometers averaged over each time-frequency range of interest, and the solution of the forward problem for each location on a grid specified by input voxel space. Following convention, the source power in these images was normalized per participant using a separately averaged prestimulus noise period of equal duration and bandwidth (66). MEG preprocessing and imaging used the Brain Electrical Source Analysis (BESA v. 6.1) software. Normalized source power was computed for the selected time-frequency bands over the entire brain volume per participant at 4.0 × 4.0 × 4.0 mm resolution. We then averaged images from each time-frequency bin of interest across all participants, time windows, and conditions (i.e., for the alpha ERD and gamma ERS individually, with each response averaged across all participants, both time windows, and both conditions), and identified the peak voxels of these responses, which corresponded to the left and right occipital cortices. We then extracted virtual sensors corresponding to the peak voxel per region. To create the virtual sensors, we applied the sensor weighting matrix derived through the forward computation to the preprocessed signal vector, which yielded a time series for the specific coordinate in source space. Note that this virtual sensor extraction was done per participant and condition individually, once the coordinates of interest (i.e., one per cluster) were known. Once these virtual sensors were extracted, they were transformed into the time-frequency domain (resolution: 0.5 Hz, 100 ms), averaged across hemispheres, and then the mean power (in % change from baseline) and peak frequency (to the nearest 0.5 Hz) were characterized by condition (i.e., Move and No-Move). Relationships between conditions as a function of time were assessed using repeated-measures ANOVAs, with time (Time 1 = before speed change vs. Time 2 = after speed change) and condition (Move vs. No-Move) as repeated-measures factors.

RESULTS

All participants were able to complete the task. A total of four participants were excluded due to artifacts in their MEG data; thus, 30 participants were included in the final analysis.

Sensor-Level Analysis

Statistical analysis of time-frequency spectrograms across all gradiometers, participants, and conditions revealed significant clusters in theta (4–8 Hz), alpha (10–16 Hz), and gamma (44–76 Hz) band oscillatory activity (P < 0.001). Briefly, a transient theta response was identified from ∼0.05 s to 0.25 s after initial stimulus onset. In addition, significant gamma ERS and alpha ERD responses began ∼0.15 s after stimulus onset and were each sustained through the speed change (and subsequent behavioral response, if applicable) until stimulus offset around 2.0 s. Figure 2 shows a group-averaged spectrogram from a posterior gradiometer collapsed across conditions. To address our hypotheses regarding conditional differences (Move vs. No-Move), we focused our analysis on visual oscillatory dynamics before and after the gabor stimulus speed change, up until around the average reaction time (i.e., 0.55 s after the change in stimulus speed or 1.55 s following stimulus onset). Thus, two significant time-frequency bins per oscillatory response were subjected to beamforming (alpha ERD: 10–16 Hz, 0.15–0.55 s and 1.15–1.55 s; gamma ERS: 44–76 Hz, 0.15–0.55 s and 1.15–1.55 s, where 1.0 s is the change in gabor stimulus speed). Note that these source images were derived for each participant and condition individually. Given the focus of the paper and our hypotheses, we did not further examine the theta response.

Figure 2.

Left: time-frequency spectrogram of peak alpha and gamma activity, with frequency (Hz) shown on the y axis and time (s) shown on the x axis. The color legend is displayed below the spectrogram and denotes percentage change from baseline power (−0.7 to -0.2 s). Time-frequency bins of significant activity (alpha: 10–16 Hz from 0.15–0.55 s and 1.15–1.55 s, 1.0 s = stimulus speed change; gamma: 44–76 Hz from 0.15–0.55 s and 1.15–1.55 s) were imaged relative to the bin’s baseline power. Right: two-dimensional (2-D) maps of the whole head sensor array are shown. Sensors with significant alpha ERD activity are denoted in cooler colors in the left array, whereas sensors of significant gamma ERS activity are denoted with warmer colors in the right array (P values < 0.001, corrected). ERD, event-related desynchronization; ERS, event-related synchronization.

Effects of Movement on Primary Visual Oscillations

Whole brain images of gamma ERS and alpha ERD activity were separately averaged across both time windows and conditions to identify the spatial locations of each response for subsequent virtual sensor analysis. For the alpha ERD, this revealed a strong bilateral desynchronization in the left and right lateral occipital cortices (Fig. 3). The gamma ERS response was found to be more focal and constrained to the left and right medial occipital cortices (Fig. 3). To quantify potential differences in the production of visual oscillatory activity during movement, we extracted virtual sensors (i.e., voxel time series) for each condition from each of these peak voxel coordinates. We had no hypotheses regarding the potential laterality of condition effects, so we averaged virtual sensor data across the left and right hemispheres, and then derived peak frequency and mean power metrics for each oscillatory response per time window and condition. These metrics were then compared using repeated measures ANOVAs with time (Time 1 = before speed change, vs. Time 2 = after speed change) and condition (Move vs. No-Move) as repeated-measures factors (as described in methods, MEG Imaging, Virtual Sensor Extraction, and Statistics). Note that to calculate peak frequency with adequate precision, we had to resample our voxel time series (i.e., 0.5 Hz, 100 ms resolution).

Figure 3.

Top: the group-averaged whole brain map (pseudo t) depicting gamma ERS activity across both conditions and time windows is displayed centrally. Plots of power (left, in % change from baseline) and peak frequency (right; in Hz) for each condition (Red: Move, Blue: No-Move) are shown for Time 1 (i.e., before speed change) and Time 2 (after speed change). There was a main effect of time on gamma power (P = 0.034), such that power decreased from Time 1 to Time 2. However, there was no effect of condition, nor was there a time-by-condition interaction (both P values > 0.05). There were no effects of time or condition, nor was there a time-by-condition interaction on peak gamma frequency (all P values > 0.05). Bottom: the group-averaged whole brain map (pseudo t) depicting alpha ERD activity across both conditions and time windows is displayed centrally. Box-and-whisker plots of power (left, in % change from baseline) and peak frequency (right; in Hz) for each condition are shown. There were significant main effects of time and condition on alpha ERD, as well as a significant time-by-condition interaction, such that the alpha ERD became stronger (i.e., more negative from baseline) from Time 1 to Time 2 in the Move relative to the No-Move condition (P = 0.017). There was also a significant effect of time on alpha peak frequency, such that peak frequency significantly decreased from Time 1 to Time 2 across conditions (P = 0.001). There was no main effect of condition, nor was there a time-by-condition interaction on alpha peak frequency (both P values > 0.05). ERD, event-related desynchronization; ERS, event-related synchronization.

For alpha ERD power, there was a main effect of time, F(1,29) = 63.37, P < 0.001, partial η² = 0.686, such that alpha ERD power became stronger (more negative) from Time 1 to Time 2. There was also a main effect of condition, F(1,29) = 5.06, P = 0.032, partial η² = 0.148, such that alpha ERD was stronger overall in the Move relative to the No-Move condition. Finally, there was a time-by-condition interaction, F(1,29) = 6.43, P = 0.017, partial η² = 0.181, such that the increase in alpha ERD power (i.e., greater difference from baseline) with time was larger for the Move condition relative to the No-Move condition. For alpha ERD frequency, there was a significant main effect of time, F(1,29) = 12.38, P = 0.001, partial η² = 0.299, such that there was a decrease in alpha ERD frequency from Time 1 to Time 2. On the contrary, there was no main effect of condition, F(1,29) = 1.78, P = 0.193, partial η² = 0.058, nor was there a significant time-by-condition interaction, F(1,29) = 0.14, P = 0.711, partial η² = 0.005. For gamma power, there was a significant effect of time, F(1,29) = 4.96, P = 0.034, partial η² = 0.146, such that gamma ERS power decreased from Time 1 to Time 2. However, there was no main effect of condition, F(1,29) = 0.31, P = 0.583, partial η² = 0.011, nor was there a time-by-condition interaction, F(1,29) = 0.22, P = 0.641, partial η² = 0.008. There were no significant effects of or interactions between time and condition on gamma ERS frequency, time: F(1,29) = 0.36, P = 0.552, partial η² = 0.012; condition: F(1,29) = 0.06, P = 0.804, partial η² = 0.002; time-by-condition: F(1,29) = 0.002, P = 0.963, partial η² < 0.001. Power and frequency characteristics for the alpha ERD and gamma ERS are shown for each time window and condition in Fig. 3.

Although we found significant differences in the alpha ERD power from Time 1 to Time 2 between the two conditions (i.e., Move vs. No-Move), these differences could be due to the movement itself or to heightened attention to the behaviorally relevant stimulus. In other words, the stronger alpha ERD (i.e., more negative from baseline) during the Move condition in Time 2 could be due to attention or to the movement itself. To help clarify whether these differences were due to attention or movement, we performed follow-up analyses looking at the temporal progression of the alpha ERD response after the speed change, with the hypothesis that if differences in alpha ERD power were due to heightened attention, then conditional differences would peak at the presentation of the speed change and be sustained or dissipate as the participant got closer to the average reaction time. On the contrary, if the differences were predominantly due to the movement itself, the conditional difference in alpha ERD power would increase closer to the average reaction time. To this end, we performed an additional 2 × 6 repeated-measures ANOVA with condition (Move vs. No-Move) and time (in 0.1 s increments from 1.0 s to 1.6 s; 6 windows) as within-subjects variables. Mauchly’s test of sphericity was significant for the time factor, thus Greenhouse–Geisser correction was applied to our reporting of the main effect of time, as well as the time-by-condition interaction. As expected, we found a main effect of condition, F(1,29) = 6.11, P = 0.020, partial η² = 0.174. In addition, there was a main effect of time, F(1.16, 33.66) = 8.64, P = 0.004, partial η² = 0.230 such that the alpha ERD became stronger (i.e., more negative) as a function of time across conditions. There was also a trending time-by-condition interaction following Greenhouse–Geisser correction, F(1.19, 34.61) = 3.00, P = 0.086 (uncorrected P = 0.013), partial η² = 0.094. Post hoc testing showed a trending linear effect (P = 0.078, partial η² = 0.103), such that the conditional difference in alpha ERD power became linearly larger with time, suggesting that the conditional effects seen in our original analysis are likely due to the movement and not attention to the stimulus. Nonetheless, this analysis was only trending so these results should be interpreted with caution. Timeseries of alpha ERD and gamma ERS activity are shown in Fig. 4.

Figure 4.

Left: virtual sensor time series of the alpha ERD response, with % change from baseline on the y axis and time in seconds (s) on the x axis and each condition plotted separately (Red: Move, Blue: No-Move). The dotted white lines denote stimulus changes (i.e., stimulus onset, speed change, stimulus offset), whereas the gray line denotes the average reaction time. Differences in alpha ERD as a function of time and condition were compared from the speed change to the average reaction time in 0.1 s time windows using a 2 × 6 repeated-measures ANOVA. There was a main effect of time on the alpha ERD (P = 0.004), such that the response became stronger (more negative) with time. There was also a significant main effect of condition (P = 0.020), such that the alpha ERD was stronger in the Move condition relative to the No-Move condition. There was a trending time-by-condition interaction, where the difference in alpha ERD as a function of condition became larger with time (P = 0.086). Right: virtual sensor time series of gamma ERS power (in % change from baseline, y axis) as a function of time (in s; x axis) for each condition (Red: Move, Blue: No-Move). The dotted white lines denote changes in the stimulus (i.e., stimulus onset, speed change, stimulus offset), whereas the gray line denotes the average reaction time. There were no significant conditional differences in gamma ERS power, so this data were not subjected to follow-up analyses. ERD, event-related desynchronization; ERS, event-related synchronization.

DISCUSSION

The goal of the current study was to investigate whether movement alters the oscillatory dynamics that serve visual processing. Consistent with previous studies (1, 47–49), the gabor visual stimuli used in our experiment elicited temporally overlapping and robust gamma ERS and alpha ERD activity in the occipital cortices. In line with our hypotheses, alpha ERD peak frequency was lower and the magnitude of the response was larger (i.e., more negative from baseline) in the Move condition relative to the No-Move condition. However, surprisingly movement did not significantly alter gamma ERS peak frequency or power. For the rest of the DISCUSSION we discuss the implications of these findings for the development of future visual tasks.

As stated, the alpha ERD power was shown to be significantly stronger (i.e., more negative relative to baseline) in the Move condition relative to the No-Move condition, and this difference increased with time from the stimulus change to the average reaction time. This pattern of results suggests that alpha ERD activity is robustly enhanced in the presence of a concurrent movement. Prior work has shown that alpha activity in both occipital and precentral regions is important for visuomotor integration (52, 53), thus it is possible that this enhanced depression of occipital alpha during movement serves to aid in this process. Future work should more precisely investigate how alpha oscillatory dynamics relate to this visuomotor integration. Although alpha activity is paramount to basic visual processing, previous studies have already implicated alpha oscillations with a number of high-order cognitive processes (6–14, 34, 37, 38). Other work have shown the critical role of occipital alpha ERD responses in attentional processing (6–14). Nonetheless, our follow-up analysis of how the strength of the alpha ERD evolves differently with time based on the Move/No-Move condition suggests that the attentional demands were not entirely driving the observed effects. Briefly, if differences in alpha ERD power between Move and No-Move were due to attentional processing, one would expect differences to peak early after the stimulus speed change. On the contrary, our analysis showed that differences between the conditions grew as the participants got closer to moving, and then the differences dissipated after the average reaction time. It is possible that the demands of the task were not sufficient to induce attention-related differences, whereas differences due to movement were more pronounced. However, this analysis was only trending, so follow up investigations should more precisely parse apart the impact of attentional processing and movement by modulating both the relevance of the cue and the movement parameters. Nevertheless, this is the first study, in our knowledge, to systematically investigate whether responding to a simple visual event impacts alpha occipital dynamics, and the results suggest that requiring a response enhances alpha oscillations. These results are informative for future visual paradigm design and integration.

Interestingly, we found a significant decrease in gamma ERS power from Time 1 to Time 2, but no effects of condition nor a time-by-condition interaction. There were no significant effects of time or condition on gamma peak frequency. This is particularly surprising, given the importance of gamma oscillations in both movement (17, 19, 20, 67, 68), attention (6–11, 13), and visual processing (47–49). Indeed, our findings suggest a decrease in the strength of the gamma ERS from Time 1 to Time 2, which is the opposite of many attention-related studies of gamma responses (6, 9). We posit that this discrepancy in findings is due to differences in the task instructions in our study. That is, the speed change in our task was behaviorally relevant in both conditions, as it determined whether the participant would move or not, but the stimulus speed itself was not meaningful, and thus the underlying oscillatory dynamics serving visual processing may not be impacted. Thus, while occipital gamma oscillatory activity is uniquely sensitive to complex visual stimulus parameters, such as the size or shape of the stimulus, it may be relatively immune to certain other characteristics of visual paradigms unrelated to the stimulus itself, such as the presence or absence of a behavioral response. Of course, future work should investigate whether other task parameters impact occipital gamma oscillatory activity. For example, recent work suggests that gamma oscillatory activity is significantly modulated by high-order cognitive processes, such as motor conflict (18, 67), working memory (16, 38), attention (6–11, 13), and executive function (69). Therefore, as with alpha, future studies should further discriminate the potential interaction between oscillatory gamma serving cognitive functions and that of the subsequent behavioral response. Nevertheless, taken together, this pattern of results further supports the notion that alpha and gamma oscillatory activity serve divergent aspects of basic visual processing in the context of movement.

In summary, the presence of a motor response to a visual stimulus significantly strengthens alpha ERD responses in the primary visual cortices, but does not appear to impact visual gamma power or alpha or gamma peak frequency. Such a relationship is not surprising, given the integration of these two systems in the service of visuomotor control (54). We suggest that this information should be considered in the design of new visual experiments meant to elicit these responses. Specifically, given that the presence of a movement does not significantly modulate gamma oscillatory activity but increases the strength of the alpha ERD, investigators should continue to implement a behavioral response during or after a visual stimulus. This not only enhances the alpha ERD response, but also ensures that the participant is paying attention to the stimulus.

Finally, we would like to acknowledge some limitations of the current study. First, the task design was such that potential attentional differences between conditions were not easily tested. Although our follow-up analyses preliminarily suggest that our differences were due to movement and not attention, it is possible that with more participants or a task that elicited a longer reaction time, the conditional relationship between alpha activity and time could turn out to be asymptotic rather than linear (as qualitatively suggested by Fig. 4, though the statistical relationship itself was linear in nature). Such a finding would not definitively support either an attention- or movement-related effect. It is important to note that a group-averaged time series by condition does not perfectly reflect the within-subject comparisons made in the time-by-condition follow-up ANOVA, but nonetheless, this apparent discrepancy should be explored. Alternatively, it is possible that the attentional component of our task was not sufficient to illuminate attention-related fluctuations in alpha activity. Future studies should more directly modulate attention-related visuomotor interactions with regards to visual oscillatory dynamics. In addition, while our significant effects were considered “strong” by most standards (70), multiple-comparisons correction for the four ANOVAs would make the significant condition-by-time interaction in alpha power only trending (P = 0.066). Although two of these ANOVAs were planned and directly probed our hypotheses, the other two were exploratory in nature. Thus, follow-up studies should be performed to confirm that alpha ERD power is uniquely and significantly impacted by movement.

Despite these limitations, our analyses of the evolution of the alpha ERD response suggest that differences were due to the movement and not to attention. Future studies should fully address the interaction between movement production and attention allocation to better parse apart these components in the context of visual oscillatory activity. Conversely, future research should also investigate the role of visual gamma and alpha oscillatory activity on motor dynamics. That said, the current study is the first, in our knowledge, to systematically investigate the impact of movement on visual oscillatory dynamics and to provide helpful information to be used in the implementation of visual task paradigms.

DATA AVAILABILITY

Data will be available upon request in writing to the corresponding author.

GRANTS

This work was supported by National Institutes of Health Grants P20GM144641 (to E.H.-G. and T.W.W.), R01MH116782 (to T.W.W.), R01MH118013 (to T.W.W.), R01MH121101 (to T.W.W.), F31AG055332 (to A.I.W.), and F32NS119375 (to A.I.W.).

DISCLAIMERS

The content is solely the responsibility of the authors and does not reflect the official views of the National Institutes of Health (NIH).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

E.H.-G., A.I.W., M.D.S., and T.W.W. conceived and designed research; E.H.-G., M.D.S., and J.A.E. performed experiments; E.H.-G., A.I.W., C.M.E., T.R.J., and T.W.W. analyzed data; E.H.-G., A.I.W., C.M.E., M.D.S., T.R.J., J.A.E., and T.W.W. interpreted results of experiments; E.H.-G., A.I.W., and T.R.J. prepared figures; E.H.-G., A.I.W., C.M.E., T.R.J., and T.W.W. drafted manuscript; E.H.-G., A.I.W., C.M.E., M.D.S., T.R.J., J.A.E., and T.W.W. edited and revised manuscript; E.H.-G., A.I.W., C.M.E., M.D.S., T.R.J., J.A.E., and T.W.W. approved final version of manuscript.