- Email: [email protected]
Alpha Oscillations Serve to Protect Working Memory Maintenance against Anticipated Distracters
Current Biology 22, 1969–1974, October 23, 2012 ª2012 Elsevier Ltd All rights reserved
http://dx.doi.org/10.1016/j.cub.2012.08.029
Report Alpha Oscillations Serve to Protect Working Memory Maintenance against Anticipated Distracters Mathilde Bonnefond1,* and Ole Jensen1,* 1Radboud University Nijmegen, Donders Institute for Brain, Cognition and Behaviour, Centre for Cognitive Neuroimaging, Kapittelweg 29, 6525 EN Nijmegen, The Netherlands
revealed that the reaction times for strong distracters (725 6 35 ms) was significantly longer than for weak distracters (699 6 31 ms; p < 0.05; Figure 1B). The findings suggest that strong distracters are incorporated into WM in some trials.
Summary
Occipitotemporal Alpha Power Increases Prior to the Anticipation of Strong Distracters To characterize the oscillatory activity reflecting the anticipation of distracters, we quantified the properties of the MEG data in the interval before the strong versus the weak distracters (0 < t < 1.1 s). The time-frequency representation (TFR) of power revealed larger 8–12 Hz alpha activity over posterior sensors prior to the distracter onset (Figures 2A and 2B). A cluster-based randomization test (see Supplemental Information available online) showed that the effect was significant in the alpha band in the 0.6–1.1 s interval over left-lateralized parieto-occipital sensors (p < 0.05; Figure 2B). The sources reflecting the alpha power increase were located in the occipital cortex extending into the left temporal cortex (Figure 2C).
When operating in a complex world, it is essential to have mechanisms that can suppress distracting information [1, 2]. Such mechanisms might be related to neuronal oscillations, which are known to be involved in gating of incoming information [3]. We here apply a working memory (WM) task to investigate how neuronal oscillations are involved in the suppression of distracting information that can be predicted in time. We used a modified Sternberg WM task in which distracters were presented in the retention interval, while we recorded the ongoing brain activity using magnetoencephalography. The data revealed a robust adjustment of the phase of alpha oscillations in anticipation of the distracter. In trials with strong phase adjustment, response times to the memory probe were reduced. Further, the power of alpha oscillations increased prior to the distracter and predicted performance. Our findings demonstrate that the doors of perception close when a distracter is expected. The phase adjustment of the alpha rhythm adds to the computational versatility of brain oscillations, because such a mechanism allows for modulating neuronal processing on a fine temporal scale. Results To investigate the role of oscillatory activity in distracter suppression in working memory (WM) tasks, we designed a modified version of the Sternberg paradigm. A memory set of four consonants was presented to the subject followed by a probe. In the retention interval, we presented either weak (a symbol) or strong (a letter) distracters in different blocks such that the subjects could anticipate the nature of the distracter (Figure 1A). To optimize the chances for identifying a phase effect in the 10 Hz band, we flashed the distracters for 33 ms. Importantly, the onset of the distracter at t = 1.1 s was predictable across trials. The ongoing brain activity was recorded using magnetoencephalography (MEG) from 18 subjects performing this task. Reaction Times Longer for Strong than Weak Distracters Trials It is well established that the reaction time to the probe increases with memory load in the Sternberg task [4]. Thus the potential incorporation of a distracter in WM should be reflected in longer reaction times. A two-way repeatedmeasures ANOVA analysis with the factors ‘‘distracter’’ (weak versus strong) and ‘‘probe type’’ (old versus new)
*Correspondence: [email protected] (M.B.), ole.jensen@ donders.ru.nl (O.J.)
An Increase in Alpha Activity Protects against Anticipated Distracters We set out to determine whether alpha power is associated with distracter suppression as assessed by the reaction times to the probe. We computed the average alpha power (8 Hz < f < 12 Hz; 0.6 < t < 1.1 s) for the cluster of sensors shown in Figure 2B for fast and slow reaction times. This was done by sorting the trials according to reaction times in five bins. Trials from the second and fourth bin were considered fast and slow in order to ignore trials with either too long or too short reaction times because they are likely to be dominated by premature button presses or lapses in attention (see also Supplemental Information). A two-way repeatedmeasures ANOVA with factors ‘‘distracter type’’ and ‘‘reaction time’’ revealed a significant interaction between reaction time and distracter type (p < 0.05). A post hoc tests showed that alpha power was significantly higher for fast than for slow reaction times trials in the strong distracter condition (p < 0.05; Figure 2D). These results suggest that stronger alpha activity in anticipation of a strong distracter leads to a better distracter suppression. Alpha Power Predicts Individual Differences in Performance We performed an analysis comparing individual abilities to suppress distracters in relation to the ability to adjust the alpha activity prior to a distracter. Over subjects we correlated the difference in reaction time and the difference in alpha power (data derived from left occipital-temporal cortex using a beamformer spatial filter; see Supplemental Information) between weak and strong distracters. We observed a negative correlation: the stronger the difference in power between anticipated distracter types, the smaller the difference in reaction times (r = 20.56; p = 0.02; Figure 2E). These data suggest that individuals with a better ability to modulate their alpha activity are less impaired by strong distracters.
Current Biology Vol 22 No 20 1970
Figure 1. The Modified Sternberg Paradigm (A) A set of four consonants was presented sequentially at a rate of 1,133 ms per letter. As in the conventional Sternberg paradigm, subjects indicated by button press whether the probe was part of the memory set (‘‘old’’ or ‘‘new’’). In the retention interval, either a ‘‘weak’’ (a symbol) or ‘‘strong’’ (a consonant) distracter was presented. The distracter strength could be anticipated by the subjects because weak and strong distracters were grouped in blocks of trials. (B) The hit rate and reactions times in response to the memory probe (error bars represent the SEM). Although the hit rate was independent of distracter type, the response times were significantly longer for the strong compared to weak distracters. This indicates that strong distracters are encoded in working memory in some trials and interferes with retrieval.
Alpha Oscillations Align in Phase Prior to Anticipated Strong Distracters Anticipatory phase adjustment could provide a mechanism for gain regulation in association with temporal expectation [3]. We aimed to determine whether alpha oscillations can adjust in phase to optimally suppress the distracter. To this end, we computed the phase-locking factor (PLF) over trials for each condition. The TFR of the PLF (thresholded at 0.087; see Supplemental Information) over the cluster of sensors shown Figure 2B revealed phase locking in the alpha band prior to the distracter onset (Figure 3A). The topography of the PLF (8 Hz < f < 12 Hz; 0.6 < t < 1.1 s) considering all sensors revealed a left-lateralized distribution that matched the power difference between conditions (cf. Figures 2B and Figures 3B). Further analyses revealed that the phase effects are not trivially explained by systematic variations in power (see Supplemental Information). Next, we considered the difference in PLF for strong versus weak distracters (Figure 3C) (the PLF was transformed to Rayleigh Z values to control for biases due to differences in trial numbers; see Supplemental Information). The difference in phase locking increased in anticipation of the strong distracter (Figure 3C). This excludes
Figure 2. The Power of Alpha Band Activity Prior to the Distracter Onset (A) Time-frequency representation of power when comparing strong and weak distracters. The alpha power was higher for strong compared to weak distracters. The arrow indicates the onset of the distracter. (B) The topography of the alpha power (8–12 Hz; 0.6–1.1 s; combined planar gradient) when comparing strong to weak distracters. The cluster randomization analysis revealed a significant cluster (marked by asterisks; p < 0.05) in posterior sensors. (C) The source analysis of the alpha activity (8–12 Hz; 0.6–1.1 s) for strong compared to weak distracters revealed the strongest contributions from left occipitotemporal cortices (MNI coordinates: [252 262 26]; t values are masked corresponding to a p value of 0.05). To further support this result, we also compared predistracter period (0.6 s < t < 1.1 s) to preprobe period (1.833 < t < 2.233 s). We found overlapping sources is occipitoparietal cortex. (D) The difference in alpha power (0.6 < t < 1.1s; 8 < f < 12 Hz) when comparing trials with fast and slow reaction times for, respectively, weak and strong distracters (error bars represent the SEM). The log-ratio of alpha power was significantly larger for strong compared to weak distracters. (E) Differences in alpha power (strong compared to weak distracters) correlated negatively with differences in reaction times (strong compared to weak distracters) over subjects, i.e., subjects in which the alpha power increased in anticipation of a strong distracter were also subjects that were less slowed down during memory probing.
that the phase locking was explained by sustained alpha oscillations elicited by the last memory item. The difference was significant when averaged over the 0.6–1.1 s interval (p < 0.01; Figure 3D). These findings demonstrate that alpha phase is adjusted in anticipation of a predictable strong distracter. To further substantiate our findings, we investigated the phase adjustment using a spatial filter aimed at extracting the activity from the occipital-temporal source. Using a linearly
Alpha Oscillations Protect Working Memory 1971
Figure 4. Assessing Difference in Predistracter Phase for Slow and Fast Reaction Times
Figure 3. Phase Adjustment of Alpha Activity Prior to Distracter Onset (A) The time-frequency representation of the phase-locking factor (PLF) in the retention interval over sensors of interest (see Figure 2B for sensors selection; PLF thresholded at 0.087 corresponding to p < 0.01 in the Rayleigh test) reveals that alpha oscillations are adjusted in phase prior to the onset of the distracter. The strong PLF values observed in the theta band (5–8 Hz) a t = w0.4 s and t = w1.1 s reflect activity evoked by, respectively, the last memory item and the distracter. Note that the acausal temporal smoothing associated with the phase estimation was approximately three cycles long, i.e., 300 ms for 10 Hz and 500 ms for 6 Hz. The arrow indicates the onset of the distracter. We further observed a significant PLF in 14 of 18 subjects (0.6 < t < 1.1 s; 8 < f < 12 Hz; Rayleigh test: p < 0.01). (B) The topography of the PLF (0.6 < t < 1.1 s; 8 < f < 12 Hz) reveals the strongest values in posterior sensors. (C) Time-frequency representation of the difference in PLF (transformed to Rayleigh Z values where Z = number of trials 3 PLF2) comparing strong and weak distracters (for sensors marked by asterisks in Figure 2B). The difference is most pronounced in the alpha band just prior to the onset of the distracter; i.e., the phase adjustment is not explained by sustained oscillations elicited by the last memory item. (D) The PLF is significantly higher in the strong than in the weak distracter condition over the sensors of interest (p< 0.01; error bars represent the SEM). (E) The time-locked responses (event-related fields) prior to the distracter. The single trials from the left occipital-temporal source were derived using a beam forming approach. The time-locked responses were averaged over subjects. A clear modulation in the alpha band is seen prior to the distracter (as indicated by small green arrows). a.u., arbitrary units. See also Figure S1.
constrained minimum variance beamformer [5], we obtained the time course for each trial from seven grid points in the left occipital-temporal cortex (see Figure 2C). The phase adjustment should as well be observable in the trial average just prior to distracter onset. The grand average over trials and subjects is shown in Figure 3E (baseline: 20.233 < t < 20.033 s, low-pass filtered at 30 Hz). The phase-adjusted
(A) The difference in phase between fast and slow trials was quantified using a phase-difference index (error bars represent the SEM). We found a significant phase difference index for the strong but not the weak distracter trials. (B) The time-locked responses (event-related fields) prior to the distracter according to trials with fast (in red) and slow (in blue) reaction times at source level. A clear lag can be observed for three cycles of alpha activity prior to the distracter (as indicated by small green arrows). a.u. = arbitrary units. See also Figure S2.
alpha activity prior to the distracter was clearly apparent in this averaged trace (marked by green arrows). Note that the phase-adjusted alpha activity occurred just prior to the distracter and cannot be explained by sustained oscillations elicited by the last memory item at t = 0 s. We also found evidence for frontal sources in which the phase adjustment of alpha activity was more pronounced just prior to strong compared to weak distracters (see Supplemental Information). In future research, it would be of interest to investigate whether these frontal regions are involved in exercising a top-down drive that adjusts the alpha activity in posterior regions ([6, 7]; see Supplemental Information; Figure S1A). Alpha Phase Predicts the Efficiency of the Suppression of the Processing of the Distracter We further tested whether the phase-locking values for slow and fast trials were associated with a difference in absolute alpha phases. This was done using a phase-difference index [8, 9]. The phase difference index in Figure 4A shows the difference in PLFs when comparing true versus randomized trials for fast versus slow reaction times trials. For the strong distracter, the significant phase difference index suggests that fast and slow distracter trials are associated with different phases of the alpha activity at the onset of the distracter. Figure 4B shows the averaged traces for fast and slow reaction times trials for data from the occipitotemporal source in the strong distracter condition. This analysis confirms that fast and slow reaction time trials are associated with different alpha phases. Our key interpretation is that if the distracter arrives at a certain alpha phase, it will be more intrusive
Current Biology Vol 22 No 20 1972
as compared to the reverse phase. Further, the results demonstrate that the anticipatory phase adjustment has functional implications, and that the phase adjustment is not trivially a consequence of the power increase prior to the distracter. Also note the slow modulation of the ERFs in Figure 3E. This modulation is reflected in the delta band but is similar for high and low distracter trials (see Figure S2). This could suggest that the phase of the delta rhythm is involved in controlling the temporal modulation, but not magnitude, of the alpha oscillations. We conclude that anticipatory phase adjustments mainly in the alpha band serve to protect WM against distracters. Unpredictable Compared to Predictable Distracter Onsets Result in Longer Reaction Times If the temporal predictability of the distracter allows for suppression by a phase-adjustment of the alpha oscillations, then a predistracter delay varying by w100 ms (matching the duration of an alpha cycle) should reduce performance. In a behavioral study conducted outside the MEG system (see Supplemental Information), we compared fixed (1.1.s) to variable delays (from 1.03 to 1.16 s; 33 ms steps) between the last memory item and the distracter. The total duration of the retention interval was kept identical for both conditions. Fixed intervals resulted in significantly shorter reaction times (664 6 30 ms versus 687 6 33 ms; 12 participants; p < 0.05). These data suggest that when subjects can precisely anticipate the distracter they are better at suppressing it. Discussion The present experiment was designed to identify the mechanism responsible for suppressing anticipated intruding information. In a task where subjects could anticipate the strength and exact timing of distracters, we observed both a stronger power increase and phase adjustment of alpha activity in occipitotemporal areas prior to strong compared to weak distracter onsets. Importantly, we demonstrated that trials with fast and slow reaction times were associated with different absolute predistracter alpha phases and a stronger alpha power in the faster trials. These findings suggest that a failure to adjust the alpha phase and power just prior to a distracter will disturb WM processes. These results underscore that alpha activity is responsible for active functional inhibition of sensory regions [10, 11]. For instance, previous studies have shown that when attention is allocated to the right hemifield, the alpha activity decreases over the left visual system while it increases relatively in the right hemisphere. The same holds for the somatosensory system [12–21]. These modulations in alpha activity are predictive of performance [15, 18]. Decreases in alpha activity have also been shown to correlate with temporal expectations [22, 23]. We here show for the first time that alpha is modulated according to the strength of distracters and that alpha power is linked to the suppression of distracting information (see also [24]). The source of the alpha activity included the left fusiform gyrus, an area known to be engaged in letter processing [25–27]. Moreover, the activity in this area predicted individual differences in reaction time between the strong and the weak distracter conditions. This result indicates that distracter suppression in particular relies on the inhibition of representational specific areas. Further it extends results from a previous study showing that subjects with a stronger posterior alpha
increase during the retention period in the Sternberg task also perform faster [28]. The finding that alpha phase adjusts prior to the distracter is highly novel. Clearly, the alpha activity is rhythmic and the phase of the oscillatory activity has been shown to modulate processing [8, 29–33]. We now show that participants can adjust alpha phase when the timing of a distracter is predictable. Such an adjustment of oscillatory activity has been shown in attentional tasks in monkeys with respect to delta oscillations [3]. In this study, the monkeys had to detect a target either in the visual or the auditory domain both presented in alternation at 1.5 Hz. The phase of the delta activity (w1.5 Hz) was top-down adjusted in the visual cortex for the attended visual stimulus (visual-attention condition) to appear at a high-excitability phase and the unattended visual stimulus (auditory-attention condition) to appear at a low-excitability phase. Our findings extend this concept: in the Lakatos et al. study [3], the frequency of the oscillations was determined by the cadence of the task. We show that an intrinsically generated cortical rhythm can adjust in phase. Further, this alpha phase adjustment is consistent with the behavioral experiment showing that distracters predictable in time are associated with better performance compared to distracters presented with a w100 ms jitter. One might suspect that the phase adjustment just prior to the distracter is explained by alpha activity evoked by the last memory, which sustained in phase until the appearance of the distracter. This concern is ruled out by the observation that the behaviorally relevant alpha adjustment increases just prior to the distracter onset (see Figures 3E and Figures 4B; Figure S1B). Our findings demonstrating alpha phase adjustment imply that the human brain is able to predict events with a temporal precision on the order of 20–30 ms or less. This is consistent with various findings demonstrating that humans are able to discriminate fine temporal durations [34–41]. We further showed that alpha phase adjustment is predictive of fast and slow reaction times to the probe. We propose that a trial in which performance is poor is a result of an insufficient adjustment of alpha phase possibly due to lapses in attention. Although the duration of the distracter was 33 ms, we expect that the results would be reproduced when using longer lasting distracters because it is the stimulus onset that defines the critical moment for when our attention is drawn. From a mechanistic point of view, one might ask why an adjustment of both alpha magnitude and phase is required for optimal performance. It has been proposed that the alpha rhythm is a consequence of pulses of inhibition that can vary in magnitude [10, 11, 42, 43]. This notion has recently received support from monkey recordings showing that alpha oscillations inhibit neuronal firing selectively at specific phases [44]. An adjustment in alpha magnitude only would provide a suboptimal mechanism for functional inhibition, because information appearing between the inhibitory pulses will not be suppressed; thus, there is a need for alpha phase adjustment as well. Our findings are consistent with the notion that a key function of the central nervous system is to predict upcoming events [45]. We here provide insight to the physiological substrate of temporal predictions. In this study, we demonstrated that alpha phase can be adjusted in anticipation of an event in order to close the doors of perception when intruding information might disturb ongoing processes. This new principle adds considerable to the versatility of the alpha rhythm in the context neuronal
Alpha Oscillations Protect Working Memory 1973
processing. The stage is now set for investigating how general this principle is [46].
17.
Supplemental Information Supplemental Information includes two figures, Supplemental Results and Discussion, and Supplemental Experimental Procedures and can be found with this article online at http://dx.doi.org/10.1016/j.cub.2012.08.029.
18.
19. Acknowledgments 20. The authors gratefully acknowledge The Netherlands Organization for Scientific Research (NWO) VICI grant number 453-09-002, ‘‘The healthy brain’’ funded by the Netherlands Initiative Brain and Cognition, a part of the Organization for Scientific Research (NWO) under grant number 056-14-011 and the Fyssen funding scheme. We thank Paul Gaalman for assistance with structural MRI collection and Sander Berends for assistance with MEG recordings. Received: January 31, 2012 Revised: July 21, 2012 Accepted: August 15, 2012 Published online: October 4, 2012
21.
22.
23.
24. References 1. Zanto, T.P., and Gazzaley, A. (2009). Neural suppression of irrelevant information underlies optimal working memory performance. J. Neurosci. 29, 3059–3066. 2. Chadick, J.Z., and Gazzaley, A. (2011). Differential coupling of visual cortex with default or frontal-parietal network based on goals. Nat. Neurosci. 14, 830–832. 3. Lakatos, P., Karmos, G., Mehta, A.D., Ulbert, I., and Schroeder, C.E. (2008). Entrainment of neuronal oscillations as a mechanism of attentional selection. Science 320, 110–113. 4. Sternberg, S. (1966). High-speed scanning in human memory. Science 153, 652–654. 5. Van Veen, B.D., van Drongelen, W., Yuchtman, M., and Suzuki, A. (1997). Localization of brain electrical activity via linearly constrained minimum variance spatial filtering. IEEE Trans. Biomed. Eng. 44, 867–880. 6. Siegel, M., Donner, T.H., and Engel, A.K. (2012). Spectral fingerprints of large-scale neuronal interactions. Nat. Rev. Neurosci. 13, 121–134. 7. Zanto, T.P., Rubens, M.T., Thangavel, A., and Gazzaley, A. (2011). Causal role of the prefrontal cortex in top-down modulation of visual processing and working memory. Nat. Neurosci. 14, 656–661. 8. Dugue´, L., Marque, P., and VanRullen, R. (2011). The phase of ongoing oscillations mediates the causal relation between brain excitation and visual perception. J. Neurosci. 31, 11889–11893. 9. Vanrullen, R., Busch, N.A., Drewes, J., and Dubois, J. (2011). Ongoing EEG phase as a trial-by-trial predictor of perceptual and attentional variability. Front Psychol 2, 60. 10. Jensen, O., and Mazaheri, A. (2010). Shaping functional architecture by oscillatory alpha activity: gating by inhibition. Front Hum. Neurosci. 4, 186. 11. Klimesch, W., Sauseng, P., and Hanslmayr, S. (2007). EEG alpha oscillations: the inhibition-timing hypothesis. Brain Res. Brain Res. Rev. 53, 63–88. 12. Rihs, T.A., Michel, C.M., and Thut, G. (2007). Mechanisms of selective inhibition in visual spatial attention are indexed by alpha-band EEG synchronization. Eur. J. Neurosci. 25, 603–610. 13. Rihs, T.A., Michel, C.M., and Thut, G. (2009). A bias for posterior alphaband power suppression versus enhancement during shifting versus maintenance of spatial attention. Neuroimage 44, 190–199. 14. Sauseng, P., Klimesch, W., Heise, K.F., Gruber, W.R., Holz, E., Karim, A.A., Glennon, M., Gerloff, C., Birbaumer, N., and Hummel, F.C. (2009). Brain oscillatory substrates of visual short-term memory capacity. Curr. Biol. 19, 1846–1852. 15. Thut, G., Nietzel, A., Brandt, S.A., and Pascual-Leone, A. (2006). Alphaband electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts visual target detection. J. Neurosci. 26, 9494–9502. 16. Van Der Werf, J., Jensen, O., Fries, P., and Medendorp, W.P. (2008). Gamma-band activity in human posterior parietal cortex encodes the
25.
26.
27.
28.
29. 30.
31.
32. 33.
34. 35. 36.
37. 38. 39. 40. 41.
motor goal during delayed prosaccades and antisaccades. J. Neurosci. 28, 8397–8405. Worden, M.S., Foxe, J.J., Wang, N., and Simpson, G.V. (2000). Anticipatory biasing of visuospatial attention indexed by retinotopically specific alpha-band electroencephalography increases over occipital cortex. J. Neurosci. 20, RC63. Haegens, S., Ha¨ndel, B.F., and Jensen, O. (2011). Top-down controlled alpha band activity in somatosensory areas determines behavioral performance in a discrimination task. J. Neurosci. 31, 5197–5204. Hari, R., and Salmelin, R. (1997). Human cortical oscillations: a neuromagnetic view through the skull. Trends Neurosci. 20, 44–49. van Ede, F., Jensen, O., and Maris, E. (2010). Tactile expectation modulates pre-stimulus beta-band oscillations in human sensorimotor cortex. Neuroimage 51, 867–876. van Ede, F., de Lange, F., Jensen, O., and Maris, E. (2011). Orienting attention to an upcoming tactile event involves a spatially and temporally specific modulation of sensorimotor alpha- and beta-band oscillations. J. Neurosci. 31, 2016–2024. Rohenkohl, G., and Nobre, A.C. (2011). a oscillations related to anticipatory attention follow temporal expectations. J. Neurosci. 31, 14076– 14084. Lima, B., Singer, W., and Neuenschwander, S. (2011). Gamma responses correlate with temporal expectation in monkey primary visual cortex. J. Neurosci. 31, 15919–15931. Haegens, S., Luther, L., and Jensen, O. (2011). Somatosensory anticipatory alpha activity increases to suppress distracting input. J. Cogn. Neurosci. 24, 677–685. James, K.H., James, T.W., Jobard, G., Wong, A.C., and Gauthier, I. (2005). Letter processing in the visual system: different activation patterns for single letters and strings. Cogn. Affect. Behav. Neurosci. 5, 452–466. Joseph, J.E., Gathers, A.D., and Piper, G.A. (2003). Shared and dissociated cortical regions for object and letter processing. Brain Res. Cogn. Brain Res. 17, 56–67. Longcamp, M., Anton, J.L., Roth, M., and Velay, J.L. (2003). Visual presentation of single letters activates a premotor area involved in writing. Neuroimage 19, 1492–1500. Scheeringa, R., Petersson, K.M., Oostenveld, R., Norris, D.G., Hagoort, P., and Bastiaansen, M.C. (2009). Trial-by-trial coupling between EEG and BOLD identifies networks related to alpha and theta EEG power increases during working memory maintenance. Neuroimage 44, 1224–1238. Busch, N.A., Dubois, J., and VanRullen, R. (2009). The phase of ongoing EEG oscillations predicts visual perception. J. Neurosci. 29, 7869–7876. Mathewson, K.E., Gratton, G., Fabiani, M., Beck, D.M., and Ro, T. (2009). To see or not to see: prestimulus alpha phase predicts visual awareness. J. Neurosci. 29, 2725–2732. Scheeringa, R., Mazaheri, A., Bojak, I., Norris, D.G., and Kleinschmidt, A. (2011). Modulation of visually evoked cortical FMRI responses by phase of ongoing occipital alpha oscillations. J. Neurosci. 31, 3813–3820. Vanrullen, R., and Dubois, J. (2011). The psychophysics of brain rhythms. Front. Psychol. 2, 203. Romei, V., Gross, J., and Thut, G. (2012). Sounds reset rhythms of visual cortex and corresponding human visual perception. Curr. Biol. 22, 807–813. Nobre, A.C., and O’Reilly, J. (2004). Time is of the essence. Trends Cogn. Sci. 8, 387–389. Nobre, A., Correa, A., and Coull, J. (2007). The hazards of time. Curr. Opin. Neurobiol. 17, 465–470. McAuley, J.D., and Jones, M.R. (2003). Modeling effects of rhythmic context on perceived duration: a comparison of interval and entrainment approaches to short-interval timing. J. Exp. Psychol. Hum. Percept. Perform. 29, 1102–1125. Barnes, R., and Jones, M.R. (2000). Expectancy, attention, and time. Cognit. Psychol. 41, 254–311. Allan, L.G., and Kristofferson, A. (1974). Judgments about duration of brief stimuli. Percept. Psychophys. 15, 434–440. Friberg, A., and Sundberg, J. (1995). Time discrimination in a monotonic, isochronous sequence. J. Acoust. Soc. Am. 98, 2524–2531. Repp, B.H. (2000). Compensation for subliminal timing perturbations in perceptual-motor synchronization. Psychol. Res. 63, 106–128. Praamstra, P., Turgeon, M., Hesse, C.W., Wing, A.M., and Perryer, L. (2003). Neurophysiological correlates of error correction in sensorimotor-synchronization. Neuroimage 20, 1283–1297.
Current Biology Vol 22 No 20 1974
42. Jensen, O., van Dijk, H., and Mazaheri, A. (2010). Amplitude asymmetry as a mechanism for the generation of slow evoked responses. Clin. Neurophysiol. 121, 1148–1149, author reply 1149–1150. 43. Jensen, O., Bonnefond, M., and VanRullen, R. (2012). An oscillatory mechanism for prioritizing salient unattended stimuli. Trends Cogn. Sci. 16, 200–206. 44. Haegens, S., Na´cher, V., Luna, R., Romo, R., and Jensen, O. (2011). a-Oscillations in the monkey sensorimotor network influence discrimination performance by rhythmical inhibition of neuronal spiking. Proc. Natl. Acad. Sci. USA 108, 19377–19382. 45. Friston, K. (2010). The free-energy principle: a unified brain theory? Nat. Rev. Neurosci. 11, 127–138. 46. Fell, J., and Axmacher, N. (2011). The role of phase synchronization in memory processes. Nat. Rev. Neurosci. 12, 105–118.
http://dx.doi.org/10.1016/j.cub.2012.08.029
Report Alpha Oscillations Serve to Protect Working Memory Maintenance against Anticipated Distracters Mathilde Bonnefond1,* and Ole Jensen1,* 1Radboud University Nijmegen, Donders Institute for Brain, Cognition and Behaviour, Centre for Cognitive Neuroimaging, Kapittelweg 29, 6525 EN Nijmegen, The Netherlands
revealed that the reaction times for strong distracters (725 6 35 ms) was significantly longer than for weak distracters (699 6 31 ms; p < 0.05; Figure 1B). The findings suggest that strong distracters are incorporated into WM in some trials.
Summary
Occipitotemporal Alpha Power Increases Prior to the Anticipation of Strong Distracters To characterize the oscillatory activity reflecting the anticipation of distracters, we quantified the properties of the MEG data in the interval before the strong versus the weak distracters (0 < t < 1.1 s). The time-frequency representation (TFR) of power revealed larger 8–12 Hz alpha activity over posterior sensors prior to the distracter onset (Figures 2A and 2B). A cluster-based randomization test (see Supplemental Information available online) showed that the effect was significant in the alpha band in the 0.6–1.1 s interval over left-lateralized parieto-occipital sensors (p < 0.05; Figure 2B). The sources reflecting the alpha power increase were located in the occipital cortex extending into the left temporal cortex (Figure 2C).
When operating in a complex world, it is essential to have mechanisms that can suppress distracting information [1, 2]. Such mechanisms might be related to neuronal oscillations, which are known to be involved in gating of incoming information [3]. We here apply a working memory (WM) task to investigate how neuronal oscillations are involved in the suppression of distracting information that can be predicted in time. We used a modified Sternberg WM task in which distracters were presented in the retention interval, while we recorded the ongoing brain activity using magnetoencephalography. The data revealed a robust adjustment of the phase of alpha oscillations in anticipation of the distracter. In trials with strong phase adjustment, response times to the memory probe were reduced. Further, the power of alpha oscillations increased prior to the distracter and predicted performance. Our findings demonstrate that the doors of perception close when a distracter is expected. The phase adjustment of the alpha rhythm adds to the computational versatility of brain oscillations, because such a mechanism allows for modulating neuronal processing on a fine temporal scale. Results To investigate the role of oscillatory activity in distracter suppression in working memory (WM) tasks, we designed a modified version of the Sternberg paradigm. A memory set of four consonants was presented to the subject followed by a probe. In the retention interval, we presented either weak (a symbol) or strong (a letter) distracters in different blocks such that the subjects could anticipate the nature of the distracter (Figure 1A). To optimize the chances for identifying a phase effect in the 10 Hz band, we flashed the distracters for 33 ms. Importantly, the onset of the distracter at t = 1.1 s was predictable across trials. The ongoing brain activity was recorded using magnetoencephalography (MEG) from 18 subjects performing this task. Reaction Times Longer for Strong than Weak Distracters Trials It is well established that the reaction time to the probe increases with memory load in the Sternberg task [4]. Thus the potential incorporation of a distracter in WM should be reflected in longer reaction times. A two-way repeatedmeasures ANOVA analysis with the factors ‘‘distracter’’ (weak versus strong) and ‘‘probe type’’ (old versus new)
*Correspondence: [email protected] (M.B.), ole.jensen@ donders.ru.nl (O.J.)
An Increase in Alpha Activity Protects against Anticipated Distracters We set out to determine whether alpha power is associated with distracter suppression as assessed by the reaction times to the probe. We computed the average alpha power (8 Hz < f < 12 Hz; 0.6 < t < 1.1 s) for the cluster of sensors shown in Figure 2B for fast and slow reaction times. This was done by sorting the trials according to reaction times in five bins. Trials from the second and fourth bin were considered fast and slow in order to ignore trials with either too long or too short reaction times because they are likely to be dominated by premature button presses or lapses in attention (see also Supplemental Information). A two-way repeatedmeasures ANOVA with factors ‘‘distracter type’’ and ‘‘reaction time’’ revealed a significant interaction between reaction time and distracter type (p < 0.05). A post hoc tests showed that alpha power was significantly higher for fast than for slow reaction times trials in the strong distracter condition (p < 0.05; Figure 2D). These results suggest that stronger alpha activity in anticipation of a strong distracter leads to a better distracter suppression. Alpha Power Predicts Individual Differences in Performance We performed an analysis comparing individual abilities to suppress distracters in relation to the ability to adjust the alpha activity prior to a distracter. Over subjects we correlated the difference in reaction time and the difference in alpha power (data derived from left occipital-temporal cortex using a beamformer spatial filter; see Supplemental Information) between weak and strong distracters. We observed a negative correlation: the stronger the difference in power between anticipated distracter types, the smaller the difference in reaction times (r = 20.56; p = 0.02; Figure 2E). These data suggest that individuals with a better ability to modulate their alpha activity are less impaired by strong distracters.
Current Biology Vol 22 No 20 1970
Figure 1. The Modified Sternberg Paradigm (A) A set of four consonants was presented sequentially at a rate of 1,133 ms per letter. As in the conventional Sternberg paradigm, subjects indicated by button press whether the probe was part of the memory set (‘‘old’’ or ‘‘new’’). In the retention interval, either a ‘‘weak’’ (a symbol) or ‘‘strong’’ (a consonant) distracter was presented. The distracter strength could be anticipated by the subjects because weak and strong distracters were grouped in blocks of trials. (B) The hit rate and reactions times in response to the memory probe (error bars represent the SEM). Although the hit rate was independent of distracter type, the response times were significantly longer for the strong compared to weak distracters. This indicates that strong distracters are encoded in working memory in some trials and interferes with retrieval.
Alpha Oscillations Align in Phase Prior to Anticipated Strong Distracters Anticipatory phase adjustment could provide a mechanism for gain regulation in association with temporal expectation [3]. We aimed to determine whether alpha oscillations can adjust in phase to optimally suppress the distracter. To this end, we computed the phase-locking factor (PLF) over trials for each condition. The TFR of the PLF (thresholded at 0.087; see Supplemental Information) over the cluster of sensors shown Figure 2B revealed phase locking in the alpha band prior to the distracter onset (Figure 3A). The topography of the PLF (8 Hz < f < 12 Hz; 0.6 < t < 1.1 s) considering all sensors revealed a left-lateralized distribution that matched the power difference between conditions (cf. Figures 2B and Figures 3B). Further analyses revealed that the phase effects are not trivially explained by systematic variations in power (see Supplemental Information). Next, we considered the difference in PLF for strong versus weak distracters (Figure 3C) (the PLF was transformed to Rayleigh Z values to control for biases due to differences in trial numbers; see Supplemental Information). The difference in phase locking increased in anticipation of the strong distracter (Figure 3C). This excludes
Figure 2. The Power of Alpha Band Activity Prior to the Distracter Onset (A) Time-frequency representation of power when comparing strong and weak distracters. The alpha power was higher for strong compared to weak distracters. The arrow indicates the onset of the distracter. (B) The topography of the alpha power (8–12 Hz; 0.6–1.1 s; combined planar gradient) when comparing strong to weak distracters. The cluster randomization analysis revealed a significant cluster (marked by asterisks; p < 0.05) in posterior sensors. (C) The source analysis of the alpha activity (8–12 Hz; 0.6–1.1 s) for strong compared to weak distracters revealed the strongest contributions from left occipitotemporal cortices (MNI coordinates: [252 262 26]; t values are masked corresponding to a p value of 0.05). To further support this result, we also compared predistracter period (0.6 s < t < 1.1 s) to preprobe period (1.833 < t < 2.233 s). We found overlapping sources is occipitoparietal cortex. (D) The difference in alpha power (0.6 < t < 1.1s; 8 < f < 12 Hz) when comparing trials with fast and slow reaction times for, respectively, weak and strong distracters (error bars represent the SEM). The log-ratio of alpha power was significantly larger for strong compared to weak distracters. (E) Differences in alpha power (strong compared to weak distracters) correlated negatively with differences in reaction times (strong compared to weak distracters) over subjects, i.e., subjects in which the alpha power increased in anticipation of a strong distracter were also subjects that were less slowed down during memory probing.
that the phase locking was explained by sustained alpha oscillations elicited by the last memory item. The difference was significant when averaged over the 0.6–1.1 s interval (p < 0.01; Figure 3D). These findings demonstrate that alpha phase is adjusted in anticipation of a predictable strong distracter. To further substantiate our findings, we investigated the phase adjustment using a spatial filter aimed at extracting the activity from the occipital-temporal source. Using a linearly
Alpha Oscillations Protect Working Memory 1971
Figure 4. Assessing Difference in Predistracter Phase for Slow and Fast Reaction Times
Figure 3. Phase Adjustment of Alpha Activity Prior to Distracter Onset (A) The time-frequency representation of the phase-locking factor (PLF) in the retention interval over sensors of interest (see Figure 2B for sensors selection; PLF thresholded at 0.087 corresponding to p < 0.01 in the Rayleigh test) reveals that alpha oscillations are adjusted in phase prior to the onset of the distracter. The strong PLF values observed in the theta band (5–8 Hz) a t = w0.4 s and t = w1.1 s reflect activity evoked by, respectively, the last memory item and the distracter. Note that the acausal temporal smoothing associated with the phase estimation was approximately three cycles long, i.e., 300 ms for 10 Hz and 500 ms for 6 Hz. The arrow indicates the onset of the distracter. We further observed a significant PLF in 14 of 18 subjects (0.6 < t < 1.1 s; 8 < f < 12 Hz; Rayleigh test: p < 0.01). (B) The topography of the PLF (0.6 < t < 1.1 s; 8 < f < 12 Hz) reveals the strongest values in posterior sensors. (C) Time-frequency representation of the difference in PLF (transformed to Rayleigh Z values where Z = number of trials 3 PLF2) comparing strong and weak distracters (for sensors marked by asterisks in Figure 2B). The difference is most pronounced in the alpha band just prior to the onset of the distracter; i.e., the phase adjustment is not explained by sustained oscillations elicited by the last memory item. (D) The PLF is significantly higher in the strong than in the weak distracter condition over the sensors of interest (p< 0.01; error bars represent the SEM). (E) The time-locked responses (event-related fields) prior to the distracter. The single trials from the left occipital-temporal source were derived using a beam forming approach. The time-locked responses were averaged over subjects. A clear modulation in the alpha band is seen prior to the distracter (as indicated by small green arrows). a.u., arbitrary units. See also Figure S1.
constrained minimum variance beamformer [5], we obtained the time course for each trial from seven grid points in the left occipital-temporal cortex (see Figure 2C). The phase adjustment should as well be observable in the trial average just prior to distracter onset. The grand average over trials and subjects is shown in Figure 3E (baseline: 20.233 < t < 20.033 s, low-pass filtered at 30 Hz). The phase-adjusted
(A) The difference in phase between fast and slow trials was quantified using a phase-difference index (error bars represent the SEM). We found a significant phase difference index for the strong but not the weak distracter trials. (B) The time-locked responses (event-related fields) prior to the distracter according to trials with fast (in red) and slow (in blue) reaction times at source level. A clear lag can be observed for three cycles of alpha activity prior to the distracter (as indicated by small green arrows). a.u. = arbitrary units. See also Figure S2.
alpha activity prior to the distracter was clearly apparent in this averaged trace (marked by green arrows). Note that the phase-adjusted alpha activity occurred just prior to the distracter and cannot be explained by sustained oscillations elicited by the last memory item at t = 0 s. We also found evidence for frontal sources in which the phase adjustment of alpha activity was more pronounced just prior to strong compared to weak distracters (see Supplemental Information). In future research, it would be of interest to investigate whether these frontal regions are involved in exercising a top-down drive that adjusts the alpha activity in posterior regions ([6, 7]; see Supplemental Information; Figure S1A). Alpha Phase Predicts the Efficiency of the Suppression of the Processing of the Distracter We further tested whether the phase-locking values for slow and fast trials were associated with a difference in absolute alpha phases. This was done using a phase-difference index [8, 9]. The phase difference index in Figure 4A shows the difference in PLFs when comparing true versus randomized trials for fast versus slow reaction times trials. For the strong distracter, the significant phase difference index suggests that fast and slow distracter trials are associated with different phases of the alpha activity at the onset of the distracter. Figure 4B shows the averaged traces for fast and slow reaction times trials for data from the occipitotemporal source in the strong distracter condition. This analysis confirms that fast and slow reaction time trials are associated with different alpha phases. Our key interpretation is that if the distracter arrives at a certain alpha phase, it will be more intrusive
Current Biology Vol 22 No 20 1972
as compared to the reverse phase. Further, the results demonstrate that the anticipatory phase adjustment has functional implications, and that the phase adjustment is not trivially a consequence of the power increase prior to the distracter. Also note the slow modulation of the ERFs in Figure 3E. This modulation is reflected in the delta band but is similar for high and low distracter trials (see Figure S2). This could suggest that the phase of the delta rhythm is involved in controlling the temporal modulation, but not magnitude, of the alpha oscillations. We conclude that anticipatory phase adjustments mainly in the alpha band serve to protect WM against distracters. Unpredictable Compared to Predictable Distracter Onsets Result in Longer Reaction Times If the temporal predictability of the distracter allows for suppression by a phase-adjustment of the alpha oscillations, then a predistracter delay varying by w100 ms (matching the duration of an alpha cycle) should reduce performance. In a behavioral study conducted outside the MEG system (see Supplemental Information), we compared fixed (1.1.s) to variable delays (from 1.03 to 1.16 s; 33 ms steps) between the last memory item and the distracter. The total duration of the retention interval was kept identical for both conditions. Fixed intervals resulted in significantly shorter reaction times (664 6 30 ms versus 687 6 33 ms; 12 participants; p < 0.05). These data suggest that when subjects can precisely anticipate the distracter they are better at suppressing it. Discussion The present experiment was designed to identify the mechanism responsible for suppressing anticipated intruding information. In a task where subjects could anticipate the strength and exact timing of distracters, we observed both a stronger power increase and phase adjustment of alpha activity in occipitotemporal areas prior to strong compared to weak distracter onsets. Importantly, we demonstrated that trials with fast and slow reaction times were associated with different absolute predistracter alpha phases and a stronger alpha power in the faster trials. These findings suggest that a failure to adjust the alpha phase and power just prior to a distracter will disturb WM processes. These results underscore that alpha activity is responsible for active functional inhibition of sensory regions [10, 11]. For instance, previous studies have shown that when attention is allocated to the right hemifield, the alpha activity decreases over the left visual system while it increases relatively in the right hemisphere. The same holds for the somatosensory system [12–21]. These modulations in alpha activity are predictive of performance [15, 18]. Decreases in alpha activity have also been shown to correlate with temporal expectations [22, 23]. We here show for the first time that alpha is modulated according to the strength of distracters and that alpha power is linked to the suppression of distracting information (see also [24]). The source of the alpha activity included the left fusiform gyrus, an area known to be engaged in letter processing [25–27]. Moreover, the activity in this area predicted individual differences in reaction time between the strong and the weak distracter conditions. This result indicates that distracter suppression in particular relies on the inhibition of representational specific areas. Further it extends results from a previous study showing that subjects with a stronger posterior alpha
increase during the retention period in the Sternberg task also perform faster [28]. The finding that alpha phase adjusts prior to the distracter is highly novel. Clearly, the alpha activity is rhythmic and the phase of the oscillatory activity has been shown to modulate processing [8, 29–33]. We now show that participants can adjust alpha phase when the timing of a distracter is predictable. Such an adjustment of oscillatory activity has been shown in attentional tasks in monkeys with respect to delta oscillations [3]. In this study, the monkeys had to detect a target either in the visual or the auditory domain both presented in alternation at 1.5 Hz. The phase of the delta activity (w1.5 Hz) was top-down adjusted in the visual cortex for the attended visual stimulus (visual-attention condition) to appear at a high-excitability phase and the unattended visual stimulus (auditory-attention condition) to appear at a low-excitability phase. Our findings extend this concept: in the Lakatos et al. study [3], the frequency of the oscillations was determined by the cadence of the task. We show that an intrinsically generated cortical rhythm can adjust in phase. Further, this alpha phase adjustment is consistent with the behavioral experiment showing that distracters predictable in time are associated with better performance compared to distracters presented with a w100 ms jitter. One might suspect that the phase adjustment just prior to the distracter is explained by alpha activity evoked by the last memory, which sustained in phase until the appearance of the distracter. This concern is ruled out by the observation that the behaviorally relevant alpha adjustment increases just prior to the distracter onset (see Figures 3E and Figures 4B; Figure S1B). Our findings demonstrating alpha phase adjustment imply that the human brain is able to predict events with a temporal precision on the order of 20–30 ms or less. This is consistent with various findings demonstrating that humans are able to discriminate fine temporal durations [34–41]. We further showed that alpha phase adjustment is predictive of fast and slow reaction times to the probe. We propose that a trial in which performance is poor is a result of an insufficient adjustment of alpha phase possibly due to lapses in attention. Although the duration of the distracter was 33 ms, we expect that the results would be reproduced when using longer lasting distracters because it is the stimulus onset that defines the critical moment for when our attention is drawn. From a mechanistic point of view, one might ask why an adjustment of both alpha magnitude and phase is required for optimal performance. It has been proposed that the alpha rhythm is a consequence of pulses of inhibition that can vary in magnitude [10, 11, 42, 43]. This notion has recently received support from monkey recordings showing that alpha oscillations inhibit neuronal firing selectively at specific phases [44]. An adjustment in alpha magnitude only would provide a suboptimal mechanism for functional inhibition, because information appearing between the inhibitory pulses will not be suppressed; thus, there is a need for alpha phase adjustment as well. Our findings are consistent with the notion that a key function of the central nervous system is to predict upcoming events [45]. We here provide insight to the physiological substrate of temporal predictions. In this study, we demonstrated that alpha phase can be adjusted in anticipation of an event in order to close the doors of perception when intruding information might disturb ongoing processes. This new principle adds considerable to the versatility of the alpha rhythm in the context neuronal
Alpha Oscillations Protect Working Memory 1973
processing. The stage is now set for investigating how general this principle is [46].
17.
Supplemental Information Supplemental Information includes two figures, Supplemental Results and Discussion, and Supplemental Experimental Procedures and can be found with this article online at http://dx.doi.org/10.1016/j.cub.2012.08.029.
18.
19. Acknowledgments 20. The authors gratefully acknowledge The Netherlands Organization for Scientific Research (NWO) VICI grant number 453-09-002, ‘‘The healthy brain’’ funded by the Netherlands Initiative Brain and Cognition, a part of the Organization for Scientific Research (NWO) under grant number 056-14-011 and the Fyssen funding scheme. We thank Paul Gaalman for assistance with structural MRI collection and Sander Berends for assistance with MEG recordings. Received: January 31, 2012 Revised: July 21, 2012 Accepted: August 15, 2012 Published online: October 4, 2012
21.
22.
23.
24. References 1. Zanto, T.P., and Gazzaley, A. (2009). Neural suppression of irrelevant information underlies optimal working memory performance. J. Neurosci. 29, 3059–3066. 2. Chadick, J.Z., and Gazzaley, A. (2011). Differential coupling of visual cortex with default or frontal-parietal network based on goals. Nat. Neurosci. 14, 830–832. 3. Lakatos, P., Karmos, G., Mehta, A.D., Ulbert, I., and Schroeder, C.E. (2008). Entrainment of neuronal oscillations as a mechanism of attentional selection. Science 320, 110–113. 4. Sternberg, S. (1966). High-speed scanning in human memory. Science 153, 652–654. 5. Van Veen, B.D., van Drongelen, W., Yuchtman, M., and Suzuki, A. (1997). Localization of brain electrical activity via linearly constrained minimum variance spatial filtering. IEEE Trans. Biomed. Eng. 44, 867–880. 6. Siegel, M., Donner, T.H., and Engel, A.K. (2012). Spectral fingerprints of large-scale neuronal interactions. Nat. Rev. Neurosci. 13, 121–134. 7. Zanto, T.P., Rubens, M.T., Thangavel, A., and Gazzaley, A. (2011). Causal role of the prefrontal cortex in top-down modulation of visual processing and working memory. Nat. Neurosci. 14, 656–661. 8. Dugue´, L., Marque, P., and VanRullen, R. (2011). The phase of ongoing oscillations mediates the causal relation between brain excitation and visual perception. J. Neurosci. 31, 11889–11893. 9. Vanrullen, R., Busch, N.A., Drewes, J., and Dubois, J. (2011). Ongoing EEG phase as a trial-by-trial predictor of perceptual and attentional variability. Front Psychol 2, 60. 10. Jensen, O., and Mazaheri, A. (2010). Shaping functional architecture by oscillatory alpha activity: gating by inhibition. Front Hum. Neurosci. 4, 186. 11. Klimesch, W., Sauseng, P., and Hanslmayr, S. (2007). EEG alpha oscillations: the inhibition-timing hypothesis. Brain Res. Brain Res. Rev. 53, 63–88. 12. Rihs, T.A., Michel, C.M., and Thut, G. (2007). Mechanisms of selective inhibition in visual spatial attention are indexed by alpha-band EEG synchronization. Eur. J. Neurosci. 25, 603–610. 13. Rihs, T.A., Michel, C.M., and Thut, G. (2009). A bias for posterior alphaband power suppression versus enhancement during shifting versus maintenance of spatial attention. Neuroimage 44, 190–199. 14. Sauseng, P., Klimesch, W., Heise, K.F., Gruber, W.R., Holz, E., Karim, A.A., Glennon, M., Gerloff, C., Birbaumer, N., and Hummel, F.C. (2009). Brain oscillatory substrates of visual short-term memory capacity. Curr. Biol. 19, 1846–1852. 15. Thut, G., Nietzel, A., Brandt, S.A., and Pascual-Leone, A. (2006). Alphaband electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts visual target detection. J. Neurosci. 26, 9494–9502. 16. Van Der Werf, J., Jensen, O., Fries, P., and Medendorp, W.P. (2008). Gamma-band activity in human posterior parietal cortex encodes the
25.
26.
27.
28.
29. 30.
31.
32. 33.
34. 35. 36.
37. 38. 39. 40. 41.
motor goal during delayed prosaccades and antisaccades. J. Neurosci. 28, 8397–8405. Worden, M.S., Foxe, J.J., Wang, N., and Simpson, G.V. (2000). Anticipatory biasing of visuospatial attention indexed by retinotopically specific alpha-band electroencephalography increases over occipital cortex. J. Neurosci. 20, RC63. Haegens, S., Ha¨ndel, B.F., and Jensen, O. (2011). Top-down controlled alpha band activity in somatosensory areas determines behavioral performance in a discrimination task. J. Neurosci. 31, 5197–5204. Hari, R., and Salmelin, R. (1997). Human cortical oscillations: a neuromagnetic view through the skull. Trends Neurosci. 20, 44–49. van Ede, F., Jensen, O., and Maris, E. (2010). Tactile expectation modulates pre-stimulus beta-band oscillations in human sensorimotor cortex. Neuroimage 51, 867–876. van Ede, F., de Lange, F., Jensen, O., and Maris, E. (2011). Orienting attention to an upcoming tactile event involves a spatially and temporally specific modulation of sensorimotor alpha- and beta-band oscillations. J. Neurosci. 31, 2016–2024. Rohenkohl, G., and Nobre, A.C. (2011). a oscillations related to anticipatory attention follow temporal expectations. J. Neurosci. 31, 14076– 14084. Lima, B., Singer, W., and Neuenschwander, S. (2011). Gamma responses correlate with temporal expectation in monkey primary visual cortex. J. Neurosci. 31, 15919–15931. Haegens, S., Luther, L., and Jensen, O. (2011). Somatosensory anticipatory alpha activity increases to suppress distracting input. J. Cogn. Neurosci. 24, 677–685. James, K.H., James, T.W., Jobard, G., Wong, A.C., and Gauthier, I. (2005). Letter processing in the visual system: different activation patterns for single letters and strings. Cogn. Affect. Behav. Neurosci. 5, 452–466. Joseph, J.E., Gathers, A.D., and Piper, G.A. (2003). Shared and dissociated cortical regions for object and letter processing. Brain Res. Cogn. Brain Res. 17, 56–67. Longcamp, M., Anton, J.L., Roth, M., and Velay, J.L. (2003). Visual presentation of single letters activates a premotor area involved in writing. Neuroimage 19, 1492–1500. Scheeringa, R., Petersson, K.M., Oostenveld, R., Norris, D.G., Hagoort, P., and Bastiaansen, M.C. (2009). Trial-by-trial coupling between EEG and BOLD identifies networks related to alpha and theta EEG power increases during working memory maintenance. Neuroimage 44, 1224–1238. Busch, N.A., Dubois, J., and VanRullen, R. (2009). The phase of ongoing EEG oscillations predicts visual perception. J. Neurosci. 29, 7869–7876. Mathewson, K.E., Gratton, G., Fabiani, M., Beck, D.M., and Ro, T. (2009). To see or not to see: prestimulus alpha phase predicts visual awareness. J. Neurosci. 29, 2725–2732. Scheeringa, R., Mazaheri, A., Bojak, I., Norris, D.G., and Kleinschmidt, A. (2011). Modulation of visually evoked cortical FMRI responses by phase of ongoing occipital alpha oscillations. J. Neurosci. 31, 3813–3820. Vanrullen, R., and Dubois, J. (2011). The psychophysics of brain rhythms. Front. Psychol. 2, 203. Romei, V., Gross, J., and Thut, G. (2012). Sounds reset rhythms of visual cortex and corresponding human visual perception. Curr. Biol. 22, 807–813. Nobre, A.C., and O’Reilly, J. (2004). Time is of the essence. Trends Cogn. Sci. 8, 387–389. Nobre, A., Correa, A., and Coull, J. (2007). The hazards of time. Curr. Opin. Neurobiol. 17, 465–470. McAuley, J.D., and Jones, M.R. (2003). Modeling effects of rhythmic context on perceived duration: a comparison of interval and entrainment approaches to short-interval timing. J. Exp. Psychol. Hum. Percept. Perform. 29, 1102–1125. Barnes, R., and Jones, M.R. (2000). Expectancy, attention, and time. Cognit. Psychol. 41, 254–311. Allan, L.G., and Kristofferson, A. (1974). Judgments about duration of brief stimuli. Percept. Psychophys. 15, 434–440. Friberg, A., and Sundberg, J. (1995). Time discrimination in a monotonic, isochronous sequence. J. Acoust. Soc. Am. 98, 2524–2531. Repp, B.H. (2000). Compensation for subliminal timing perturbations in perceptual-motor synchronization. Psychol. Res. 63, 106–128. Praamstra, P., Turgeon, M., Hesse, C.W., Wing, A.M., and Perryer, L. (2003). Neurophysiological correlates of error correction in sensorimotor-synchronization. Neuroimage 20, 1283–1297.
Current Biology Vol 22 No 20 1974
42. Jensen, O., van Dijk, H., and Mazaheri, A. (2010). Amplitude asymmetry as a mechanism for the generation of slow evoked responses. Clin. Neurophysiol. 121, 1148–1149, author reply 1149–1150. 43. Jensen, O., Bonnefond, M., and VanRullen, R. (2012). An oscillatory mechanism for prioritizing salient unattended stimuli. Trends Cogn. Sci. 16, 200–206. 44. Haegens, S., Na´cher, V., Luna, R., Romo, R., and Jensen, O. (2011). a-Oscillations in the monkey sensorimotor network influence discrimination performance by rhythmical inhibition of neuronal spiking. Proc. Natl. Acad. Sci. USA 108, 19377–19382. 45. Friston, K. (2010). The free-energy principle: a unified brain theory? Nat. Rev. Neurosci. 11, 127–138. 46. Fell, J., and Axmacher, N. (2011). The role of phase synchronization in memory processes. Nat. Rev. Neurosci. 12, 105–118.