- Email: [email protected]
Cortical dynamics of selective attention to somatosensory events
NeuroImage 49 (2010) 1777–1785
Contents lists available at ScienceDirect
NeuroImage j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / y n i m g
Cortical dynamics of selective attention to somatosensory events C. Dockstader a,b, D. Cheyne a,c,d, R. Tannock a,b,e,⁎ a
Neuroscience & Mental Health Program, The Hospital for Sick Children, 555 University Avenue, Room 4265, Toronto, Ontario, Canada M5G 1X8 Department of Psychiatry, The Hospital for Sick Children, Toronto, Ontario, Canada c Department of Diagnostic Imaging, The Hospital for Sick Children, Toronto, Ontario, Canada d Department of Medical Imaging, University of Toronto, Toronto, Ontario, Canada e Human Development and Applied Psychology, Ontario Institute for Studies in Education at the University of Toronto, Toronto, Ontario, Canada b
a r t i c l e
i n f o
Article history: Received 11 May 2009 Revised 26 August 2009 Accepted 16 September 2009 Available online 23 September 2009 Keywords: Attention Somatosensory cortex Cortical oscillations Magnetoencephalography Alpha Beta Gamma Phase-locking Median nerve stimulation
a b s t r a c t Recent studies have shown evidence of somatosensory deficits in individuals with attentional difficulties yet relatively little is known about the role of attention in the processing of somatosensory input. Neuromagnetic imaging studies have shown that rhythmic oscillations within the human somatosensory cortex are strongly modulated by somatosensory stimulation and may reflect the normal processing of such stimuli. However, few studies have examined how attention influences these cortical oscillations. We examined attentional effects on human somatosensory oscillations during median nerve stimulation by conducting time–frequency analyses of neuromagnetic recordings in healthy adults. We found that selective attention modulated somatosensory oscillations in the alpha, beta, and gamma bands that were both phaselocked and non-phase-locked to the stimulus. In the primary somatosensory cortex (SI), directing the subject's attention toward the somatosensory stimulus resulted in increased gamma band power (30–55 Hz) that was phase-locked to stimulus onset. Directed attention also produced an initial suppression (desynchrony) followed by enhancement (synchrony) of beta band power (13–25 Hz) that was not phase-locked to the stimulus. In the secondary somatosensory cortex (SII), directing attention towards the stimulus increased phase-locked alpha (7–9 Hz) power approximately 30 ms after onset of phase-locked gamma in SI, followed by a non-phase-locked increase in alpha power. We suggest that earlier phase-locked oscillatory power may reflect the relay of input from SI to SII, whereas later non-phase-locked rhythms reflect stimulus-induced oscillations that are modulated by selective attention and may thus reflect enhanced processing of the stimulus underlying the perception of somatosensory events. © 2009 Elsevier Inc. All rights reserved.
Introduction Attention to a particular object or event while ignoring others (i.e., selective attention) optimizes task performance by ensuring that the underlying neural activities operate efficiently (Fahle, 2009; Hillyard, 1993; Kok et al., 2006; Naatanen and Michie, 1979). Neural efficiency can be indexed by measuring changes in synchronous firing of large cortical assemblies within specific frequency bands. Substantial evidence from studies of auditory and visual selective attention indicates that neural synchrony increases within the gamma band (30–120 Hz) in the modality-relevant primary cortex and also in higher-order regions such as association areas and the frontal lobes (Driver and Frackowiak, 2001; Herrmann and Knight, 2001; Knudsen, 2007). Moreover, recent evidence suggests that changes in the synchrony of lower frequencies such as alpha (8–12 Hz) (Palva and Palva, 2007; Thut et al., 2006; Worden et al., 2000) and beta (13–30 Hz) (Bekisz and Wrobel, 2003; ⁎ Corresponding author. Neuroscience & Mental Health Program, The Hospital for Sick Children, 555 University Avenue, Room 4265, Toronto, Ontario, Canada M5G 1X8. Fax: +1 416 813 6565. E-mail address: [email protected] (R. Tannock). 1053-8119/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2009.09.035
Wrobel, 2000; Wrobel et al., 2007) also accompany increases in attention. The basic premise underlying the effectiveness of neural synchrony is that large groups of neurons are more likely to communicate effectively with their downstream recipients when they are firing in synchrony than when they are asynchronous, due to more efficient summation of their action potentials (Fries et al., 2001; Niebur, 2002; Niebur et al., 2002). Thus, selective attention and the potentiation of neural synchrony optimizes neural communication within and between behaviourally-related regions. A number of new findings highlight somatosensory processing deficits in individuals with attentional difficulties (Broring et al., 2008; Georgiou-Karistianis et al., 2003; Parush et al., 1997, 2007; Scherder et al., 2008). However, little is known about the effect of selective attention on somatosensory processing in either healthy individuals or clinical populations. Currently two studies of healthy adults have investigated non-painful, attention-related changes to neural synchrony in the human somatosensory cortex. Both suggest that attention directed toward a passive somatosensory event increases gamma synchrony in the somatosensory cortex (Bauer et al., 2006; Ray et al., 2008). A somatosensory stimulus elicits an early, event-related potential or field in the primary somatosensory cortex (SI) that is phase-locked
1778
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
to the onset of the somatosensory event. It is represented by a sharp, transient increase in synchrony that is tightly time-locked to the onset of an exogenous somatosensory event and is thought to be the mechanism by which a percept is formed (Engel et al., 2001; Fell et al., 2003; Varela et al., 2001). Although this response can be modulated by top-down processes, its appearance depends upon an external, stimulus-driven sensory event. A somatosensory stimulus can also change the synchrony of ongoing, endogenous oscillations. It may generate an event-related desynchronization (ERD) or event-related synchronization (ERS) of cortical rhythms within a particular frequency range. ERD and ERS are non-phase-locked responses and vary from trial to trial, occurring at slightly different times following the stimulus event. ERD and ERS can also occur prior to a known stimulus event and it is believed that these oscillations are easily influenced by cognitive processes such as attention, orienting, and anticipation (Neuper et al., 2006; Sochurkova et al., 2006; Stancak, 2006). Although both phase-locked and non-phase-locked activities reflect changes in neural synchrony, they are thought to be functionally distinct processes in neural communication. To our knowledge, there have been no studies that have specifically examined phase-locked versus non-phase-locked oscillations in SI during selective attention. Moreover, previous studies have not compared these to oscillations further downstream in the secondary somatosensory cortex (SII). Thus, the current study aimed to characterize attentional effects on both power and phase in SI and SII in adults who received passive somatosensory stimulation.
onset of the next train of stimuli. Participants were presented with 200 trains of stimuli over a 12-min experiment. Stimulus presentation was controlled by Presentation Software (Neurobehavioral Systems, Inc., Albany, CA). Stimuli were presented under the following two experimental conditions, counterbalanced for condition order. (1) Attend to MNS: participants were instructed to attend to the electrical stimuli and count the number of stimulus trains. To ensure eyes remained open and head remained still, participants were asked to focus on a yellow circle (2 in. diameter, 4 ft above MEG helmet). (b) Ignore MNS: participants were instructed to ignore the median nerve stimuli and attend to a video (“Pingu the Penguin”) that was presented, without sound, on a back-projection screen via two mirrors, on a screen 75 cm in front of the participants. To ensure attention was directed to the video during this condition, participants were asked to count the number of times the penguin trumpeted (32 events) and the number of times he caught a fish (8 events) (both were random events interspersed throughout the video). All participants reported the correct number of incidents with an error rate of ±6% indicating that they were able to comply with instructions and successfully direct attention to the designated events at a behavioural level. This study was approved by the Institutional Research Ethics Board and informed written consent was obtained for each participant. Additionally, it was in compliance with national legislation and the Code of Ethical Principles for Medical Research Involving Human Subjects of the World Medical Association (Declaration of Helsinki).
Materials and methods Data analyses Participants Twelve healthy, right-handed adults (6F) were recruited from a hospital newsletter advertisement. The mean age of the participants was 28.3 ± 3.1 SEM years. Participants were screened with a telephone-based Intake Screening Questionnaire (for psychopathology and education level). Participants were excluded if they (1) reported a current mental health disorder, (2) had a history of neurological disorders, or (3) had any head injury involving the loss of consciousness. Equipment A 151 channel MEG system (VSM MedTech Ltd., Vancouver, Canada) was used to measure somatosensory fields. Participants lay supine with their head resting in the MEG helmet in the magnetically shielded room. The MEG signals were filtered with an online bandpass of 0.3–300 Hz and recorded at a 1250 Hz sampling rate. Head position in relation to the MEG sensors was determined by measuring the magnetic field generated by three fiducial reference coils just before and after each experimental session. T1-weighted structural magnetic resonance images (MRI) (axial 3D spoiled gradient echo sequence) were obtained for each participant using a 1.5 Tesla Signal Advantage system (GE Medical Systems, Milwaukee). During MRI data acquisition, three radiographic markers were positioned at the same anatomical landmarks as the fiduciary coils to allow coregistration of the MEG and MRI data. Single equivalent current dipole (ECD) models were also fit to the N20 m median nerve responses in order to confirm coregistration accuracy.
Spatial localization of cortical oscillations was performed using a spatial filtering (beamforming) approach and time–frequency analysis applied to the time course of neural activity at the identified locations of peak activity within predetermined frequency bands. Beamformer spatial analysis Initial spatial analyses were performed based on a minimumvariance beamformer method (synthetic aperture magnetometry: SAM (Robinson and Vrba, 1999)), by computing spatial filters from the single-trial data filtered from 1–200 Hz. We used a scalar version of the minimum-variance beamformer algorithm that estimates a single optimal current orientation at each voxel (Cheyne et al., 2007; Sekihara et al., 2004). Changes in source power were computed using the SAM pseudo-t statistic (Robinson and Vrba, 1999) which involved subtracting source power in the control period prior to stimulus presentation (−200 ms to 0 ms relative to stimulus onset) from an active period (0–200 ms post-stimulus (or gap onset)). This allowed us to identify peak locations of broad-band increases in activity to the somatosensory stimulus in the SAM images for further analysis.
Paradigms
Virtual sensor analyses Based on the peak SI and SII locations as determined with SAM, we created single-trial time series of source activity or ‘virtual sensors’ by passing broad-spectrum (1–200 Hz) single-trial MEG signals through the spatial filter for each of the locations. This produced time series representing the averaged changes in source power for both phaselocked response and non-phase-locked responses of SI and SII over time.
Stimuli were non-painful, 0.2 ms pulses of electrical current, just above motor threshold (eliciting a small, passive, thumb twitch) applied cutaneously to the right median nerve. Stimuli were presented with a constant stimulus onset of 670 ms, in trains of four consecutive events followed by a pause of 1340 ms before the
Group virtual sensor analyses Virtual sensors for SI and SII locations were also averaged across participants. To determine statistical differences in amplitude and latency between conditions, we performed repeated measures ANOVAs separately for SI and SII data to examine main effects of
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
1779
Table 1 Time–frequency windows for which statistical comparisons were made. SI Gamma Alpha Beta ERD Beta ERS
Phase-locked 30–55 Hz, 20–45 ms 8–12 Hz, 0–100 ms ⁎ ⁎
SI Non-phase-locked 30–55 Hz, 20–70 ms 8–12 Hz, 160–260 ms 13–25 Hz, 100–230 ms 13–25 Hz, 370–500 ms
Phase-locked 30–55 Hz, 45–70 ms 7–9 Hz, 0–125 ms ⁎ 13–25 Hz, 480–545 ms
Non-phase-locked ⁎ 7–9 Hz, 100–350 ms ⁎ ⁎
⁎ No activity observed.
component (i.e., N20 m and P35 m) and condition (Attend to MNS/ Ignore MNS) on latency and amplitude of the virtual sensor data. We examined latencies and amplitudes of the N20 m, P35 m, P60 m, and N140 m components for SI, and N50 m, P100 m, and N200 m components for SII. Time–frequency analyses Based on the individual virtual sensors, we created time– frequency representation (TFR) plots of source power activity over a pre- and post-stimulus interval, integrated over all trials, using a Morlet wavelet transform (Tallon-Baudry et al., 1996) over three frequency bandwidths (1–30 Hz, 30–60 Hz, and 60–120 Hz). Based on the peak contralateral and ipsilateral SI locations as determined with SAM, we created single-trial time series of source activity by passing the single-trial MEG signals through the spatial filter for each of the locations to produce plots of aggregated data that displayed both stimulus phase-locked and non-phase-locked source power changes over time. Dissociating phase-locked and non-phase-locked changes To disentangle phase-locked from non-phase-locked changes we ran two additional analyses: (1) stimulus-locked averaged data were subtracted from each trial prior to computing the single-trial power, which left only the non-phase-locked components in the TFR for each of the attentional conditions, and (2) time–frequency transform was computed from the averaged data to examine only the phase-locked components in the TFR response for each condition. Group time–frequency analyses Time–frequency plots were averaged across participants. To determine statistical differences between conditions for time– frequency responses we first delineated the specific time–frequency windows for which statistical comparisons would be made. In general, for each bandwidth of interest we set the frequency
boundary to be that typically considered to be the exemplary bandwidth, as defined in this sample, and the temporal boundary to be defined by the beginning of the observed response through one complete oscillation of the bandwidth during phase-locked activities and two complete oscillations of the bandwidth during non-phaselocked activities (a temporal boundary of two frequency oscillations accounted for the vast majority of temporal jittering of the nonphase-locked responses). Time–frequency boundaries are shown in Table 1. We then applied these time–frequency boundaries to each individual's relevant TFR. Pixel values within that boundary box were averaged to obtain a single value for each individual, within each bandwidth, of each condition. Paired t-tests were performed separately for each time–frequency boundary with the factor ‘attentional condition’ (Attend to MNS, Ignore MNS) being within subjects. Results SI SI-contralateral effects of attention as a function of time Differential SAM analyses of broad-spectrum activity (1–200 Hz) in the Attend to MNS condition localized maximal activity to the thumb region of contralateral SI (SI-C) (Brodmann area 3B) during the active period for all participants (Fig. 1A). Based on this location, we created virtual sensors for each individual's SI response, for both attentional conditions. The virtual sensors shown in Fig. 1B depict the grand-averaged power changes over time of the broad-spectrum SI phase-locked response in both attentional conditions. The stimulus-related averaged signals were generally characterized by four early components: N20 m, P35 m, P60 m, and N140 m. Repeated measures ANOVA revealed a main effect of attention on amplitude in which attending to the MNS increased the overall amplitude of
Fig. 1. (A) Localization of the contralateral SI response. Example of the location of maximal SI activity during the active period following median nerve stimulation in the ‘Attend to MNS’ condition. Maximal activity occurred in the left postcentral gyrus, immediately caudal to the central sulcus (CS), in Brodmann area 3B, in every subject. Subsequent virtual sensors and TFR analyses were based upon this target location for each individual. Inset shows averaged phase-locked fields from all MEG channels for control and active periods. (B) Grand-averaged virtual sensors of the contralateral SI response. SI virtual sensors for changes in source power over the duration of the averaged trials were averaged across subjects. Inset shows the amplitude differences of the first latency peak for both conditions.
1780
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
Fig. 2. (A) Attention effects on amplitude changes in contralateral SI. Histogram representing grand-averaged changes in amplitude of the phase-locked response for the N20 m, P35 m, P60 m, and N140 m components of both Attend to MNS and Ignore MNS conditions. Overall, selective attention to the somatosensory event significantly increased the amplitude of the SI response. (B) Attention effects on latency changes in contralateral SI. Histogram representing grand-averaged changes in latency of the phase-locked response for the N20 m, P35 m, P60 m, and N140 m components of both Attend to MNS and Ignore MNS conditions. Overall, selective attention to the somatosensory event significantly reduced the latency of the SI response.
the SI virtual sensor response [F(1,11) = 38.99, p b 0.0001, ES = .78] (Fig. 2A). Overall, attention to the MNS also produced an significantly quicker virtual sensor response [F(1,11) = 9.84, p b 0.009, ES = .47] (Fig. 2B). SI-contralateral effects of attention as a function of frequency and phase In order to examine effects of somatosensory attention on the SI frequency domain, virtual sensors of selected bandwidths were created and transformed with a Morlet wavelet to produce time– frequency responses. Fig. 3A depicts the grand-averaged, phaselocked + non-phase-locked response as well as phase-locked-only and non-phase-locked-only responses. Overall, the phase-locked + non-phase-locked analysis revealed three distinct components of rhythmic changes in SI-C during both attentional conditions in response to a stimulus event: (1) an early broad-spectrum increase in power that was temporally coincident with the N20-P35-N60 phase-locked response; (2) a prolonged increase in lower band power (b13 Hz) throughout the trial; and (3) a suppression of power (eventrelated desynchrony: ERD) and subsequent rebound (event-related synchrony: ERS) in the beta band in the mid and later stages of the trial, respectively. To examine changes in power and phase with selective attention we performed statistical comparisons separately for phase-locked and non-phase-locked activities. Significant effects of attention on rhythmcity are shown in Fig. 3B. Attention to the MNS significantly increased phase-locked activity in the gamma band (30–55 Hz) centering around 20–50 ms, compared to when attention was directed away from the MNS [t(1,11) = 2.425, p b 0.05]. Moreover, selective attention to the MNS also significantly increased subsequent beta ERD and ERS activities that were non-phase-locked to stimulus onset [t(1,11) = 3.27, p b 0.01 and t(1,11) = 2.54, p b 0.05, respectively]. A moderate increase in non-phase-locked gamma activities did not reach significance [t(1,11) = 1.7, p N 0.05].
SI-ipsilateral activity Low amplitude ipsilateral SI (SI-I) activity was observed in many participants' SAM images but virtual sensor data revealed only a few detectable peaks of low amplitude (data not shown). Therefore, SI-I activity was not analyzed further. SII SII-contralateral effects of attention as a function of time Differential SAM analyses of broad-spectrum activity (1–200 Hz) in the Attend to MNS condition revealed bilateral activation of the association somatosensory cortex (SII) in the parietal operculum with stronger activation in the contralateral hemisphere (SII-C) than the ipsilateral hemisphere (SII-I) (Fig. 4A). Based on these locations, we created virtual sensors for each individual's SII-C and SII-I response, for both attentional conditions. The virtual sensors shown in Fig. 4B depict the grand-averaged power changes over time of the broadspectrum SII-C response in both attentional conditions. The stimulusrelated averaged signals were generally characterized by three components: N50 m, P100 m, and N200 m. Repeated measures ANOVA revealed a main effect of attention on SII amplitude in which attention to the MNS increased the overall power of the SII virtual sensor [F(1,11) = 22.454, p b 0.001, ES = .67] (Fig. 5A). There was no main effect of attentional condition on the latency of the SII response [F(1,11) = 3.47, p N 0.05, ES = .24] (Fig. 5B). SII-contralateral effects of attention on frequency and phase Fig. 6A depicts the grand-averaged, phase-locked + non-phaselocked response, phase-locked-only, and non-phase-locked-only responses. Overall, the phase-locked + non-phase-locked analysis showed that SII-C activity was characterized by strong, prolonged oscillations in the alpha band regardless of attentional condition. Statistical comparisons of the changes in SII-C frequency source power
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
1781
Fig. 3. (A) Attentional effects on contralateral SI oscillations. Grand-averaged contralateral SI TFRs for combined phase-locked + non-phase-locked, phase-locked-only, and nonphase-locked-only oscillations of control subjects in both attentional conditions. All plots were baselined using the average spectral energy observed in the corresponding prestimulus period (−200 to 0 ms) and represent the grand mean TFR of the individual, virtual channel, spatially-filtered single trials. Selected time–frequency boundaries are outlined on the TFR plots. Source power: A − m2 = (−n × 10−17 to +n × 10−17). (B) Statistical comparison of attentional conditions for phase-locked and non-phase-locked contralateral SI responses.
and phase between the two attentional conditions are shown in Fig. 6B. Attention to the MNS significantly increased synchrony of an early phase-locked alpha response (0–100 ms) [t(1,11) = 5.62, p b 0.001] and a subsequent non-phase-locked alpha response (100–400 ms)
[t(1,11) = 2.37, p b 0.05] compared to when the participants attended away from the somatosensory event. Some modest increases in phaselocked gamma [t(1,11) = 0.836, p N 0.05] and beta [t(1,11) = 1.12, p N 0.05] powers did not reach significance.
Fig. 4. (A) Localization of the bilateral SII response. Example of the bilateral locations of maximal SII activity during the active period following median nerve stimulation in the Attend to MNS condition. Subsequent virtual sensors and TFRs analyses were based upon these target locations. Inset shows averaged phase-locked fields from all MEG channels for control and active periods. (B) Grand-averaged virtual sensors of the contralateral SII response. Virtual sensors for changes in source power over the duration of the averaged trials were averaged across subjects.
1782
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
Fig. 5. (A) Attention effects on amplitude changes in contralateral SII. Histogram representing grand-averaged changes in amplitude of the phase-locked response for the N50 m, P100 m, and N200 m components of both Attend to MNS and Ignore MNS conditions. Overall, selective attention to the somatosensory event significantly increased the amplitude of the SII response. (B) Attention effects on latency changes in contralateral SII. Histogram representing grand-averaged changes in latency of the phase-locked response for the N50 m, P100 m, and N200 m components of both Attend to MNS and Ignore MNS conditions. Overall, selective attention to the somatosensory event did not significantly affect the latency of the SII response.
SII-ipsilateral effects of attention as a function of time Virtual sensors of the SII-I data were similar to that of SII-C although overall lower in amplitude. The first latency peak of the SII-I phase-locked response occurred at 56.2 ms ± 3.68 SEM with a peak value of 0.7 fT ± 0.43 SEM and 54.1 ms ± 7.11 SEM with a peak value of 0.79 fT ± 0.51 SEM for Attend to MNS and Ignore MNS conditions, respectively. There were no statistical differences of response time or peak values between conditions. SII-ipsilateral effects of attention on frequency and phase SII-I peaks were also characterized by strong, prolonged oscillations in the sigma band (7–9 Hz) that occurred in both attention conditions. No changes in time–frequency responses were visible in SII-I and TFR statistical analyses did not detect any differences in phase or power between conditions. Fig. 7 depicts a summary of the sequence of somatosensory oscillatory activity in SI and SII to somatosensory events and how it was influenced by attention over time. Topographical specificity of the SI and SII effects The primary visual cortex (VI) shows modulations in the alpha bandwidth in response to visual attention (Rihs et al., 2009; Fujisawa et al., 2008). Because, in our paradigms, both phase-locked and nonphase-locked SII alpha rhythms (7–9 Hz) modulated with attention we investigated participants' VI time–frequency responses to verify topographical specificity of the observed attentional effects. VI virtual sensors deriving from each participant's activation peak in the primary visual cortex were grand-averaged and transposed into TFRs. VI TFRs showed only non-phase-locked responses, consistent with the absence of repetitive, time-locked visual events. VI exhibited stronger alpha oscillations (10–12 Hz) when attention was directed to the tactile event while lower frequency oscillations (3–6 Hz) increased when participants attended to the video (and ignored the
MNS). No beta or gamma oscillations were present in either condition (see Supplementary Fig. 1). As alpha oscillations in SII and VI differed both in bandwidth and in temporal onset and offset, we propose that the somatosensory effects were highly localized and not confounded by oscillation in the visual cortex. Discussion We examined the changes in amplitude, frequency, and phase of somatosensory oscillations while manipulating selective attention. Our major findings indicate that (1) attention modulates SI phaselocked gamma rhythms and non-phase-locked beta rhythms, (2) attention modulates SII phase-locked and non-phase-locked alpha rhythms, and (3) early increases in phase-locked SI gamma precede all SII responses. Selective attention modulates power and phase of SI oscillations Gamma-band synchrony has been associated with cognitive functions such as attention, working memory, and top-down modulation of sensory processing (Fries et al., 2001; Pesaran et al., 2002). Human EEG research shows that the strength of the early phase-locked gamma response centered around 40 Hz is particularly sensitive to modulation by controlled attention (Debener et al., 2002; Fries et al., 2001; Struber et al., 2000). Our data show that selective attention to a somatosensory stimulus increased the early phaselocked SI gamma response. This increase commenced at 20 ms and was most robust around 40 Hz. Thus, the median nerve stimulation paradigm is an effective method for determining the earliest stages of attentional influence. We also compared changes in SI alpha and beta band synchrony across the two attentional conditions. In our study, SI alpha rhythms were not modulated by attention. Moreover, only alpha synchrony,
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
1783
Fig. 6. (A) Attentional effects on contralateral SII oscillations. Grand-averaged, contralateral SII time–frequency responses for combined phase-locked + non-phase-locked, phaselocked-only, and non-phase-locked-only oscillations of control subjects in both attentional conditions. Due to very low-frequency noise (2–4 Hz, throughout duration of trial), all plots were baselined using the average spectral energy observed across the trial (−670 to 0 ms). Each plot represents the grand mean TFR of the individual, virtual channel, spatiallyfiltered single trials. Selected time–frequency boundaries are outlined on the TFR plots. Source power: A − m2 = (−n × 10−17 to +n × 10−17). (B) Statistical comparison of attentional conditions for phase-locked and non-phase-locked contralateral SII responses.
but not desynchrony, was present. A large number of studies within the visual and auditory domains propose that alpha ERD reflects sensory processing while alpha ERS reflects cortical inhibition
(Sauseng et al., 2005; Worden et al., 2000; Thut et al., 2006; Rihs et al., 2007). It may be that somatosensory alpha activities may not show the same consistent relationship as the visual and auditory
Fig. 7. Summary of how attention affects phase-locked and non-phase-locked somatosensory oscillations. Electrical stimulation of the right median nerve elicits a robust, contralateral response in the left cortical hemisphere. Neuromagnetic recordings and subsequent beamformer analysis show that the two peak activities localize in contralateral SI and SII. Waveform analysis combined with time–frequency transformations indicate that the initial SI response shows amplitude peaks at 20, 35, and 60 ms following stimulus onset which coincide with phase-locked gamma activity and increase with attention to the somatosensory event. SII amplitude peaks that occur at 55 and 100 ms post-stimulus coincide with phase-locked alpha activity. All phase-locked activities are amplified by attention. Both SI and SII also show non-phase-locked oscillations in response to the somatosensory event that begin ∼100 ms after stimulus onset (beta and alpha rhythms, respectively). All non-phase-locked components are amplified with selective attention.
1784
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
systems do. Alternatively, the robust alpha synchrony we observed may be specific to the electrical stimulus applied. Correspondingly, Forss et al. (2001), Palva et al. (2005), and Dockstader et al. (2009) showed that electrical stimulation that was applied to the wrist (Forss et al.; Dockstader et al.) or the index finger (Palva et al.) elicited an immediate and robust increase in SI alpha power. This was later followed by an alpha ERD when stimulus frequency was 0.5 Hz or slower (Palva et al.; Dockstader et al.) but not when stimuli were presented rapidly (1.5–12 Hz). In our experiment (and in the case of Forss et al.) the absence of a later alpha ERD response may simply be due to the brief time period between stimulus events. Although attention did not alter SI alpha synchrony, both nonphase-locked beta ERD and ERS occurred. Sensorimotor beta oscillations have largely been associated with motor activities (Neuper and Pfurtscheller, 1996; Sochurkova et al., 2006). However, these oscillations may play a role in attention as increases in sensorimotor beta power correlate with attention directed toward a motoric event (Muthukumaraswamy and Singh, 2008) and improved sensorimotor performance (Egner and Gruzelier, 2004; Egner et al., 2004; Vernon et al., 2003). Interestingly, individuals with attention-deficit/hyperactivity disorder show diminished sensorimotor beta response to somatosensory input (Dockstader et al., 2008). Selective attention modulates power and phase of alpha (‘sigma') oscillations in SII Here we show that attention to a somatosensory event does modulate SII alpha ERS. Specifically, there was an immediate, stimulus-locked increase in 7–9 Hz synchrony when participants attended to the somatosensory event followed by a delayed and prolonged non-phase-locked increase in neural synchrony. To our knowledge, this is the first report of attention-related changes in alpha oscillations observed in human SII. This results suggests that, in SII, there is a direct and active role for alpha ERS in somatosensory perception and attention. Moreover, while the dynamics of the alpha rhythm remain unclear, this result contributes to the recent theory that cortical inhibition is only one of several roles of alpha ERS (Palva and Palva, 2007). Because SII alpha activity was restricted to a 7–9 Hz bandwidth, we suggest that the observed SII alpha activity may be related to the previously identified ‘sigma’ rhythm—an alpha rhythm variant restricted to 7–9 Hz in SII originally reported by Narici et al. (2001) and Forss et al. (2001).
We suggest that our observed sequence of phase-locked events reflects neural communication of sensory information from SI to SII that is enhanced, in both regions, with attention to the somatosensory event. Selective attention to the somatosensory stimulus also strengthened non-phase-locked SI beta ERD and ERS and SII alpha, or sigma, ERS. As the onset of non-phase-locked events began after SI and SII phase-locked activities were completed it is unlikely that the non-phase-locked oscillations reflected somatosensory stimulus perception per se. We suggest, rather, that these rhythms reflect cognitive processes driven by stimulus input and modulated by attention. Hemispheric asymmetry Attentional effects were found to be lateralized for both SI and SII, suggesting that the contralateral, rather than ipsilateral, hemisphere is involved in selective attention to somatosensory stimuli. Cortical hemispheric asymmetry has been associated with attention in healthy participants (Floel et al., 2005a; Floel et al., 2005b; Floel et al., 2001). It is also possible that lateralization effects might differ if we conducted the task again with left-handers or in right-handers while stimulating their left median nerve. However, an fMRI study showed a functional dominance of the left SII in secondary somatosensory cortex, regardless of handedness, using median nerve stimulation (Simoes and Hari, 1999). Hemispheric asymmetry may also be dependent on the particular stimulus applied as recent fMRI studies have shown that ipsilateral somatosensory responses are stronger for mechanical versus median nerve electrical stimulation (Lipton et al., 2006). Future directions Although our study revealed several novel findings, there are several remaining questions. To answer these questions, future studies may include (1) coherence analyses in which the relationship between SI gamma and SII alpha may be further defined, (2) investigating neural somatosensory responses to natural stimuli, (3) examining the relationship between enhanced neural somatosensory response and enhanced task performance, (4) determining whether early attentional changes to somatosensory input in SI and SII are independent of the priming sensory modality, and (5) applying our paradigm to individuals with attention deficits. Conclusion
Dissociating phase-locked and non-phase-locked oscillatory changes Phase-locked and non-phase-locked changes in synchrony can reflect functionally separate cognitive processes. Whereas it is suggested that phase-locked activities reflect enhanced information transfer between brain areas in response to an external event, nonphase-locked activities more likely reflect top-down influences that modify the sensitivity of a region to incoming events and (Klimesch et al., 2007; Pineda, 2005). Changes in either of these forms of activity can enhance task-related performance. For example, increases in alpha phase-locking in the visual cortex correlate with more sensitive perception (Hanslmayr et al., 2005) whereas changes in somatosensory alpha non-phase-locked activities correlate with learning (Zhuang et al., 1997) and anticipation (Babiloni et al., 2004). However, these rhythms do not appear to be critical in the perception of a somatosensory event (Dockstader et al., 2009). In our study, selective attention to a somatosensory stimulus enhanced synchrony in SI phase-locked gamma (∼20 ms) and SII phase-locked alpha ∼30 ms following the SI response. Postsynaptic targets are particularly sensitive to gamma band synchrony (Womelsdorf et al., 2006) and therefore the increase in SI gamma synchrony should indeed be followed by an increase in synchrony of the downstream SII neurons.
This is the first study to uncouple phase-locked versus non-phaselocked oscillations in SI during selective attention. It is also the first study to examine time–frequency characteristics of downstream SII activities. We suggest that early phase-locked oscillations signal the relay of perceptual input from SI to SII. Later non-phase-locked oscillations are sensitive to selective attention and may signal heightened stimulus processing. Understanding the dynamics of cortical oscillations and their relationship to attentional control will provide new insights into some of the earliest and most basic aspects of cognitive processing. Acknowledgments We thank Dr. Margot Taylor for feedback on the manuscript and Travis Mills, Sonya Bells, and Christine Popovich for their technical support. This research was supported in part by funds from a Canadian Institutes of Health Research Operating Grant (CIHR # 64279; PI is D.C.) and an operating grant from The Hospital for Sick Children Psychiatry Endowment Fun (C.D., D.C., and R.T.), as well as an Ontario Mental Health Foundation fellowship (C.D.) and the Canada Research Chairs Program (R.T.).
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.neuroimage.2009.09.035. References Babiloni, C., Brancucci, A., Arendt-Nielsen, L., Babiloni, F., Capotosto, P., Carducci, F., Cincotti, F., Romano, L., Chen, A.C., Rossini, P.M., 2004. Alpha event-related desynchronization preceding a go/no-go task: a high-resolution EEG study. Neuropsychology 18, 719–728. Bauer, M., Oostenveld, R., Peeters, M., Fries, P., 2006. Tactile spatial attention enhances gamma-band activity in somatosensory cortex and reduces low-frequency activity in parieto-occipital areas. J. Neurosci. 26, 490–501. Bekisz, M., Wrobel, A., 2003. Attention-dependent coupling between beta activities recorded in the cat's thalamic and cortical representations of the central visual field. Eur. J. Neurosci. 17, 421–426. Broring, T., Rommelse, N., Sergeant, J., Scherder, E., 2008. Sex differences in tactile defensiveness in children with ADHD and their siblings. Dev. Med. Child Neurol. 50, 129–133. Cheyne, D., Bostan, A.C., Gaetz, W., Pang, E.W., 2007. Event-related beamforming: a robust method for presurgical functional mapping using MEG. Clin. Neurophysiol. 118, 1691–1704. Debener, S., Kranczioch, C., Herrmann, C.S., Engel, A.K., 2002. Auditory novelty oddball allows reliable distinction of top-down and bottom-up processes of attention. Int. J. Psychophysiol. 46, 77–84. Dockstader, C., Gaetz, W., Cheyne, D., Wang, F., Castellanos, F.X., Tannock, R., 2008. MEG event-related desynchronization and synchronization deficits during basic somatosensory processing in individuals with ADHD. Behav. Brain Funct. 4, 8. Dockstader, C., Gaetz, W., Cheyne, D., Tannock, R., 2009. Abnormal neural reactivity to unpredictable sensory events in attention-deficit/hyperactivity disorder. Biol. Psychiatry 66, 376–383. Driver, J., Frackowiak, R.S., 2001. Neurobiological measures of human selective attention. Neuropsychologia 39, 1257–1262. Egner, T., Gruzelier, J.H., 2004. EEG biofeedback of low beta band components: frequency-specific effects on variables of attention and event-related brain potentials. Clin. Neurophysiol. 115, 131–139. Egner, T., Zech, T.F., Gruzelier, J.H., 2004. The effects of neurofeedback training on the spectral topography of the electroencephalogram. Clin. Neurophysiol. 115, 2452–2460. Engel, A.K., Fries, P., Singer, W., 2001. Dynamic predictions: oscillations and synchrony in top-down processing. Nat. Rev. Neurosci. 2, 704–716. Fahle, M., 2009. Perceptual learning and sensomotor flexibility: cortical plasticity under attentional control? Philos. Trans. R. Soc. Lond., B Biol. Sci. 364, 313–319. Fell, J., Fernandez, G., Klaver, P., Elger, C.E., Fries, P., 2003. Is synchronized neuronal gamma activity relevant for selective attention? Brain Res. Brain Res. Rev. 42, 265–272. Floel, A., Knecht, S., Lohmann, H., Deppe, M., Sommer, J., Drager, B., Ringelstein, E.B., Henningsen, H., 2001. Language and spatial attention can lateralize to the same hemisphere in healthy humans. Neurology 57, 1018–1024. Floel, A., Buyx, A., Breitenstein, C., Lohmann, H., Knecht, S., 2005a. Hemispheric lateralization of spatial attention in right- and left-hemispheric language dominance. Behav. Brain Res. 158, 269–275. Floel, A., Jansen, A., Deppe, M., Kanowski, M., Konrad, C., Sommer, J., Knecht, S., 2005b. Atypical hemispheric dominance for attention: functional MRI topography. J. Cereb. Blood. Flow Metab. 25, 1197–1208. Forss, N., Narici, L., Hari, R., 2001. Sustained activation of the human SII cortices by stimulus trains. NeuroImage 13, 497–501. Fries, P., Reynolds, J.H., Rorie, A.E., Desimone, R., 2001. Modulation of oscillatory neuronal synchronization by selective visual attention. Science 291, 1560–1563. Fujisawa, J., Touyama, H., Hirose, M., 2008. Extracting alpha band modulation during visual spatial attention without flickering stimuli using common spatial pattern. Conf. Proc. IEEE Eng. Med. Biol. Soc. 620–623. Georgiou-Karistianis, N., Howells, D., Bradshaw, J., 2003. Orienting attention in obsessive-compulsive disorder. Cogn. Behav. Neurol. 16, 68–74. Hanslmayr, S., Klimesch, W., Sauseng, P., Gruber, W., Doppelmayr, M., Freunberger, R., Pecherstorfer, T., 2005. Visual discrimination performance is related to decreased alpha amplitude but increased phase locking. Neurosci. Lett. 375, 64–68. Herrmann, C.S., Knight, R.T., 2001. Mechanisms of human attention: event-related potentials and oscillations. Neurosci. Biobehav. Rev. 25, 465–476. Hillyard, S.A., 1993. Electrical and magnetic brain recordings: contributions to cognitive neuroscience. Curr. Opin. Neurobiol. 3, 217–224. Klimesch, W., Sauseng, P., Hanslmayr, S., 2007. EEG alpha oscillations: the inhibitiontiming hypothesis. Brain Res. Rev. 53, 63–88. Knudsen, E.I., 2007. Fundamental components of attention. Annu. Rev. Neurosci. 30, 57–78. Kok, A., Ridderinkhof, K.R., Ullsperger, M., 2006. The control of attention and actions: current research and future developments. Brain Res. 1105, 1–6. Lipton, M.L., Fu, K.M., Branch, C.A., Schroeder, C.E., 2006. Ipsilateral hand input to area 3b revealed by converging hemodynamic and electrophysiological analyses in macaque monkeys. J. Neurosci. 26, 180–185. Muthukumaraswamy, S.D., Singh, K.D., 2008. Modulation of the human mirror neuron system during cognitive activity. Psychophysiology 45, 896–905.
1785
Naatanen, R., Michie, P.T., 1979. Early selective-attention effects on the evoked potential: a critical review and reinterpretation. Biol. Psychol. 8, 81–136. Narici, L., Forss, N., Jousmaki, V., Peresson, M., Hari, R., 2001. Evidence for a 7- to 9-Hz “sigma” rhythm in the human SII cortex. NeuroImage 13, 662–668. Neuper, C., Pfurtscheller, G., 1996. Post-movement synchronization of beta rhythms in the EEG over the cortical foot area in man. Neurosci. Lett. 216, 17–20. Neuper, C., Wortz, M., Pfurtscheller, G., 2006. ERD/ERS patterns reflecting sensorimotor activation and deactivation. Prog. Brain Res. 159, 211–222. Niebur, E., 2002. Electrophysiological correlates of synchronous neural activity and attention: a short review. Biosystems 67, 157–166. Niebur, E., Hsiao, S.S., Johnson, K.O., 2002. Synchrony: a neuronal mechanism for attentional selection? Curr. Opin. Neurobiol. 12, 190–194. Palva, S., Palva, J.M., 2007. New vistas for alpha-frequency band oscillations. Trends Neurosci. 30, 150–158. Palva, S., Linkenkaer-Hansen, K., Naatanen, R., Palva, J.M., 2005. Early neural correlates of conscious somatosensory perception. J. Neurosci. 25, 5248–5258. Parush, S., Sohmer, H., Steinberg, A., Kaitz, M., 1997. Somatosensory functioning in children with attention deficit hyperactivity disorder. Dev. Med. Child Neurol. 39, 464–468. Parush, S., Sohmer, H., Steinberg, A., Kaitz, M., 2007. Somatosensory function in boys with ADHD and tactile defensiveness. Physiol. Behav. 90, 553–558. Pesaran, B., Pezaris, J.S., Sahani, M., Mitra, P.P., Andersen, R.A., 2002. Temporal structure in neuronal activity during working memory in macaque parietal cortex. Nat. Neurosci. 5, 805–811. Pineda, J.A., 2005. The functional significance of mu rhythms: translating “seeing” and “hearing” into “doing”. Brain Res. Brain Res. Rev. 50, 57–68. Ray, S., Niebur, E., Hsiao, S.S., Sinai, A., Crone, N.E., 2008. High-frequency gamma activity (80–150 Hz) is increased in human cortex during selective attention. Clin. Neurophysiol. 119, 116–133. Rihs, T.A., Michel, C.M., Thut, G., 2007. Mechanisms of selective inhibition in visual spatial attention are indexed by alpha-band EEG synchronization. Eur. J. Neurosci. 25, 603–610. Rihs, T.A., Michel, C.M., Thut, G., 2009. A bias for posterior alpha-band power suppression versus enhancement during shifting versus maintenance of spatial attention. Neuroimage 44 (1), 190–199. Robinson, S., Vrba, J., 1999. Functional neuroimaging by synthetic aperture magnetometry (SAM). Tohku Univ. Press, Sendai, Japan. Sauseng, P., Klimesch, W., Stadler, W., Schabus, M., Doppelmayr, M., Hanslmayr, S., Gruber, W.R., Birbaumer, N., 2005. A shift of visual spatial attention is selectively associated with human EEG alpha activity. Eur. J. Neurosci. 22, 2917–2926. Scherder, E.J., Rommelse, N.N., Broring, T., Faraone, S.V., Sergeant, J.A., 2008. Somatosensory functioning and experienced pain in ADHD-families: a pilot study. Eur. J. Paediatr. Neurol. 12 (6), 461–469. Sekihara, K., Nagarajan, S.S., Poeppel, D., Marantz, A., 2004. Asymptotic SNR of scalar and vector minimum-variance beamformers for neuromagnetic source reconstruction. IEEE Trans. Biomed. Eng. 51, 1726–1734. Simoes, C., Hari, R., 1999. Relationship between responses to contra- and ipsilateral stimuli in the human second somatosensory cortex SII. NeuroImage 10, 408–416. Sochurkova, D., Rektor, I., Jurak, P., Stancak, A., 2006. Intracerebral recording of cortical activity related to self-paced voluntary movements: a Bereitschaftspotential and event-related desynchronization/synchronization. SEEG study. Exp. Brain Res. 173, 637–649. Stancak, A., 2006. Cortical oscillatory changes occurring during somatosensory and thermal stimulation. Prog. Brain Res. 159, 237–252. Struber, D., Basar-Eroglu, C., Hoff, E., Stadler, M., 2000. Reversal-rate dependent differences in the EEG gamma-band during multistable visual perception. Int. J. Psychophysiol. 38, 243–252. Tallon-Baudry, C., Bertrand, O., Delpuech, C., Pernier, J., 1996. Stimulus specificity of phase-locked and non-phase-locked 40 Hz visual responses in human. J. Neurosci. 16, 4240–4249. Thut, G., Nietzel, A., Brandt, S.A., Pascual-Leone, A., 2006. Alpha-band electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts visual target detection. J. Neurosci. 26, 9494–9502. Varela, F., Lachaux, J.P., Rodriguez, E., Martinerie, J., 2001. The brainweb: phase synchronization and large-scale integration. Nat. Rev. Neurosci. 2, 229–239. Vernon, D., Egner, T., Cooper, N., Compton, T., Neilands, C., Sheri, A., Gruzelier, J., 2003. The effect of training distinct neurofeedback protocols on aspects of cognitive performance. Int. J. Psychophysiol. 47, 75–85. Womelsdorf, T., Fries, P., Mitra, P.P., Desimone, R., 2006. Gamma-band synchronization in visual cortex predicts speed of change detection. Nature 439, 733–736. Worden, M.S., Foxe, J.J., Wang, N., Simpson, G.V., 2000. Anticipatory biasing of visuospatial attention indexed by retinotopically specific alpha-band electroencephalography increases over occipital cortex. J. Neurosci. 20, RC63. Wrobel, A., 2000. Beta activity: a carrier for visual attention. Acta Neurobiol. Exp. (Wars) 60, 247–260. Wrobel, A., Ghazaryan, A., Bekisz, M., Bogdan, W., Kaminski, J., 2007. Two streams of attention-dependent beta activity in the striate recipient zone of cat's lateral posterior–pulvinar complex. J. Neurosci. 27, 2230–2240. Zhuang, P., Toro, C., Grafman, J., Manganotti, P., Leocani, L., Hallett, M., 1997. Eventrelated desynchronization (ERD) in the alpha frequency during development of implicit and explicit learning. Electroencephalogr. Clin. Neurophysiol. 102, 374–381.
Contents lists available at ScienceDirect
NeuroImage j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / y n i m g
Cortical dynamics of selective attention to somatosensory events C. Dockstader a,b, D. Cheyne a,c,d, R. Tannock a,b,e,⁎ a
Neuroscience & Mental Health Program, The Hospital for Sick Children, 555 University Avenue, Room 4265, Toronto, Ontario, Canada M5G 1X8 Department of Psychiatry, The Hospital for Sick Children, Toronto, Ontario, Canada c Department of Diagnostic Imaging, The Hospital for Sick Children, Toronto, Ontario, Canada d Department of Medical Imaging, University of Toronto, Toronto, Ontario, Canada e Human Development and Applied Psychology, Ontario Institute for Studies in Education at the University of Toronto, Toronto, Ontario, Canada b
a r t i c l e
i n f o
Article history: Received 11 May 2009 Revised 26 August 2009 Accepted 16 September 2009 Available online 23 September 2009 Keywords: Attention Somatosensory cortex Cortical oscillations Magnetoencephalography Alpha Beta Gamma Phase-locking Median nerve stimulation
a b s t r a c t Recent studies have shown evidence of somatosensory deficits in individuals with attentional difficulties yet relatively little is known about the role of attention in the processing of somatosensory input. Neuromagnetic imaging studies have shown that rhythmic oscillations within the human somatosensory cortex are strongly modulated by somatosensory stimulation and may reflect the normal processing of such stimuli. However, few studies have examined how attention influences these cortical oscillations. We examined attentional effects on human somatosensory oscillations during median nerve stimulation by conducting time–frequency analyses of neuromagnetic recordings in healthy adults. We found that selective attention modulated somatosensory oscillations in the alpha, beta, and gamma bands that were both phaselocked and non-phase-locked to the stimulus. In the primary somatosensory cortex (SI), directing the subject's attention toward the somatosensory stimulus resulted in increased gamma band power (30–55 Hz) that was phase-locked to stimulus onset. Directed attention also produced an initial suppression (desynchrony) followed by enhancement (synchrony) of beta band power (13–25 Hz) that was not phase-locked to the stimulus. In the secondary somatosensory cortex (SII), directing attention towards the stimulus increased phase-locked alpha (7–9 Hz) power approximately 30 ms after onset of phase-locked gamma in SI, followed by a non-phase-locked increase in alpha power. We suggest that earlier phase-locked oscillatory power may reflect the relay of input from SI to SII, whereas later non-phase-locked rhythms reflect stimulus-induced oscillations that are modulated by selective attention and may thus reflect enhanced processing of the stimulus underlying the perception of somatosensory events. © 2009 Elsevier Inc. All rights reserved.
Introduction Attention to a particular object or event while ignoring others (i.e., selective attention) optimizes task performance by ensuring that the underlying neural activities operate efficiently (Fahle, 2009; Hillyard, 1993; Kok et al., 2006; Naatanen and Michie, 1979). Neural efficiency can be indexed by measuring changes in synchronous firing of large cortical assemblies within specific frequency bands. Substantial evidence from studies of auditory and visual selective attention indicates that neural synchrony increases within the gamma band (30–120 Hz) in the modality-relevant primary cortex and also in higher-order regions such as association areas and the frontal lobes (Driver and Frackowiak, 2001; Herrmann and Knight, 2001; Knudsen, 2007). Moreover, recent evidence suggests that changes in the synchrony of lower frequencies such as alpha (8–12 Hz) (Palva and Palva, 2007; Thut et al., 2006; Worden et al., 2000) and beta (13–30 Hz) (Bekisz and Wrobel, 2003; ⁎ Corresponding author. Neuroscience & Mental Health Program, The Hospital for Sick Children, 555 University Avenue, Room 4265, Toronto, Ontario, Canada M5G 1X8. Fax: +1 416 813 6565. E-mail address: [email protected] (R. Tannock). 1053-8119/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2009.09.035
Wrobel, 2000; Wrobel et al., 2007) also accompany increases in attention. The basic premise underlying the effectiveness of neural synchrony is that large groups of neurons are more likely to communicate effectively with their downstream recipients when they are firing in synchrony than when they are asynchronous, due to more efficient summation of their action potentials (Fries et al., 2001; Niebur, 2002; Niebur et al., 2002). Thus, selective attention and the potentiation of neural synchrony optimizes neural communication within and between behaviourally-related regions. A number of new findings highlight somatosensory processing deficits in individuals with attentional difficulties (Broring et al., 2008; Georgiou-Karistianis et al., 2003; Parush et al., 1997, 2007; Scherder et al., 2008). However, little is known about the effect of selective attention on somatosensory processing in either healthy individuals or clinical populations. Currently two studies of healthy adults have investigated non-painful, attention-related changes to neural synchrony in the human somatosensory cortex. Both suggest that attention directed toward a passive somatosensory event increases gamma synchrony in the somatosensory cortex (Bauer et al., 2006; Ray et al., 2008). A somatosensory stimulus elicits an early, event-related potential or field in the primary somatosensory cortex (SI) that is phase-locked
1778
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
to the onset of the somatosensory event. It is represented by a sharp, transient increase in synchrony that is tightly time-locked to the onset of an exogenous somatosensory event and is thought to be the mechanism by which a percept is formed (Engel et al., 2001; Fell et al., 2003; Varela et al., 2001). Although this response can be modulated by top-down processes, its appearance depends upon an external, stimulus-driven sensory event. A somatosensory stimulus can also change the synchrony of ongoing, endogenous oscillations. It may generate an event-related desynchronization (ERD) or event-related synchronization (ERS) of cortical rhythms within a particular frequency range. ERD and ERS are non-phase-locked responses and vary from trial to trial, occurring at slightly different times following the stimulus event. ERD and ERS can also occur prior to a known stimulus event and it is believed that these oscillations are easily influenced by cognitive processes such as attention, orienting, and anticipation (Neuper et al., 2006; Sochurkova et al., 2006; Stancak, 2006). Although both phase-locked and non-phase-locked activities reflect changes in neural synchrony, they are thought to be functionally distinct processes in neural communication. To our knowledge, there have been no studies that have specifically examined phase-locked versus non-phase-locked oscillations in SI during selective attention. Moreover, previous studies have not compared these to oscillations further downstream in the secondary somatosensory cortex (SII). Thus, the current study aimed to characterize attentional effects on both power and phase in SI and SII in adults who received passive somatosensory stimulation.
onset of the next train of stimuli. Participants were presented with 200 trains of stimuli over a 12-min experiment. Stimulus presentation was controlled by Presentation Software (Neurobehavioral Systems, Inc., Albany, CA). Stimuli were presented under the following two experimental conditions, counterbalanced for condition order. (1) Attend to MNS: participants were instructed to attend to the electrical stimuli and count the number of stimulus trains. To ensure eyes remained open and head remained still, participants were asked to focus on a yellow circle (2 in. diameter, 4 ft above MEG helmet). (b) Ignore MNS: participants were instructed to ignore the median nerve stimuli and attend to a video (“Pingu the Penguin”) that was presented, without sound, on a back-projection screen via two mirrors, on a screen 75 cm in front of the participants. To ensure attention was directed to the video during this condition, participants were asked to count the number of times the penguin trumpeted (32 events) and the number of times he caught a fish (8 events) (both were random events interspersed throughout the video). All participants reported the correct number of incidents with an error rate of ±6% indicating that they were able to comply with instructions and successfully direct attention to the designated events at a behavioural level. This study was approved by the Institutional Research Ethics Board and informed written consent was obtained for each participant. Additionally, it was in compliance with national legislation and the Code of Ethical Principles for Medical Research Involving Human Subjects of the World Medical Association (Declaration of Helsinki).
Materials and methods Data analyses Participants Twelve healthy, right-handed adults (6F) were recruited from a hospital newsletter advertisement. The mean age of the participants was 28.3 ± 3.1 SEM years. Participants were screened with a telephone-based Intake Screening Questionnaire (for psychopathology and education level). Participants were excluded if they (1) reported a current mental health disorder, (2) had a history of neurological disorders, or (3) had any head injury involving the loss of consciousness. Equipment A 151 channel MEG system (VSM MedTech Ltd., Vancouver, Canada) was used to measure somatosensory fields. Participants lay supine with their head resting in the MEG helmet in the magnetically shielded room. The MEG signals were filtered with an online bandpass of 0.3–300 Hz and recorded at a 1250 Hz sampling rate. Head position in relation to the MEG sensors was determined by measuring the magnetic field generated by three fiducial reference coils just before and after each experimental session. T1-weighted structural magnetic resonance images (MRI) (axial 3D spoiled gradient echo sequence) were obtained for each participant using a 1.5 Tesla Signal Advantage system (GE Medical Systems, Milwaukee). During MRI data acquisition, three radiographic markers were positioned at the same anatomical landmarks as the fiduciary coils to allow coregistration of the MEG and MRI data. Single equivalent current dipole (ECD) models were also fit to the N20 m median nerve responses in order to confirm coregistration accuracy.
Spatial localization of cortical oscillations was performed using a spatial filtering (beamforming) approach and time–frequency analysis applied to the time course of neural activity at the identified locations of peak activity within predetermined frequency bands. Beamformer spatial analysis Initial spatial analyses were performed based on a minimumvariance beamformer method (synthetic aperture magnetometry: SAM (Robinson and Vrba, 1999)), by computing spatial filters from the single-trial data filtered from 1–200 Hz. We used a scalar version of the minimum-variance beamformer algorithm that estimates a single optimal current orientation at each voxel (Cheyne et al., 2007; Sekihara et al., 2004). Changes in source power were computed using the SAM pseudo-t statistic (Robinson and Vrba, 1999) which involved subtracting source power in the control period prior to stimulus presentation (−200 ms to 0 ms relative to stimulus onset) from an active period (0–200 ms post-stimulus (or gap onset)). This allowed us to identify peak locations of broad-band increases in activity to the somatosensory stimulus in the SAM images for further analysis.
Paradigms
Virtual sensor analyses Based on the peak SI and SII locations as determined with SAM, we created single-trial time series of source activity or ‘virtual sensors’ by passing broad-spectrum (1–200 Hz) single-trial MEG signals through the spatial filter for each of the locations. This produced time series representing the averaged changes in source power for both phaselocked response and non-phase-locked responses of SI and SII over time.
Stimuli were non-painful, 0.2 ms pulses of electrical current, just above motor threshold (eliciting a small, passive, thumb twitch) applied cutaneously to the right median nerve. Stimuli were presented with a constant stimulus onset of 670 ms, in trains of four consecutive events followed by a pause of 1340 ms before the
Group virtual sensor analyses Virtual sensors for SI and SII locations were also averaged across participants. To determine statistical differences in amplitude and latency between conditions, we performed repeated measures ANOVAs separately for SI and SII data to examine main effects of
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
1779
Table 1 Time–frequency windows for which statistical comparisons were made. SI Gamma Alpha Beta ERD Beta ERS
Phase-locked 30–55 Hz, 20–45 ms 8–12 Hz, 0–100 ms ⁎ ⁎
SI Non-phase-locked 30–55 Hz, 20–70 ms 8–12 Hz, 160–260 ms 13–25 Hz, 100–230 ms 13–25 Hz, 370–500 ms
Phase-locked 30–55 Hz, 45–70 ms 7–9 Hz, 0–125 ms ⁎ 13–25 Hz, 480–545 ms
Non-phase-locked ⁎ 7–9 Hz, 100–350 ms ⁎ ⁎
⁎ No activity observed.
component (i.e., N20 m and P35 m) and condition (Attend to MNS/ Ignore MNS) on latency and amplitude of the virtual sensor data. We examined latencies and amplitudes of the N20 m, P35 m, P60 m, and N140 m components for SI, and N50 m, P100 m, and N200 m components for SII. Time–frequency analyses Based on the individual virtual sensors, we created time– frequency representation (TFR) plots of source power activity over a pre- and post-stimulus interval, integrated over all trials, using a Morlet wavelet transform (Tallon-Baudry et al., 1996) over three frequency bandwidths (1–30 Hz, 30–60 Hz, and 60–120 Hz). Based on the peak contralateral and ipsilateral SI locations as determined with SAM, we created single-trial time series of source activity by passing the single-trial MEG signals through the spatial filter for each of the locations to produce plots of aggregated data that displayed both stimulus phase-locked and non-phase-locked source power changes over time. Dissociating phase-locked and non-phase-locked changes To disentangle phase-locked from non-phase-locked changes we ran two additional analyses: (1) stimulus-locked averaged data were subtracted from each trial prior to computing the single-trial power, which left only the non-phase-locked components in the TFR for each of the attentional conditions, and (2) time–frequency transform was computed from the averaged data to examine only the phase-locked components in the TFR response for each condition. Group time–frequency analyses Time–frequency plots were averaged across participants. To determine statistical differences between conditions for time– frequency responses we first delineated the specific time–frequency windows for which statistical comparisons would be made. In general, for each bandwidth of interest we set the frequency
boundary to be that typically considered to be the exemplary bandwidth, as defined in this sample, and the temporal boundary to be defined by the beginning of the observed response through one complete oscillation of the bandwidth during phase-locked activities and two complete oscillations of the bandwidth during non-phaselocked activities (a temporal boundary of two frequency oscillations accounted for the vast majority of temporal jittering of the nonphase-locked responses). Time–frequency boundaries are shown in Table 1. We then applied these time–frequency boundaries to each individual's relevant TFR. Pixel values within that boundary box were averaged to obtain a single value for each individual, within each bandwidth, of each condition. Paired t-tests were performed separately for each time–frequency boundary with the factor ‘attentional condition’ (Attend to MNS, Ignore MNS) being within subjects. Results SI SI-contralateral effects of attention as a function of time Differential SAM analyses of broad-spectrum activity (1–200 Hz) in the Attend to MNS condition localized maximal activity to the thumb region of contralateral SI (SI-C) (Brodmann area 3B) during the active period for all participants (Fig. 1A). Based on this location, we created virtual sensors for each individual's SI response, for both attentional conditions. The virtual sensors shown in Fig. 1B depict the grand-averaged power changes over time of the broad-spectrum SI phase-locked response in both attentional conditions. The stimulus-related averaged signals were generally characterized by four early components: N20 m, P35 m, P60 m, and N140 m. Repeated measures ANOVA revealed a main effect of attention on amplitude in which attending to the MNS increased the overall amplitude of
Fig. 1. (A) Localization of the contralateral SI response. Example of the location of maximal SI activity during the active period following median nerve stimulation in the ‘Attend to MNS’ condition. Maximal activity occurred in the left postcentral gyrus, immediately caudal to the central sulcus (CS), in Brodmann area 3B, in every subject. Subsequent virtual sensors and TFR analyses were based upon this target location for each individual. Inset shows averaged phase-locked fields from all MEG channels for control and active periods. (B) Grand-averaged virtual sensors of the contralateral SI response. SI virtual sensors for changes in source power over the duration of the averaged trials were averaged across subjects. Inset shows the amplitude differences of the first latency peak for both conditions.
1780
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
Fig. 2. (A) Attention effects on amplitude changes in contralateral SI. Histogram representing grand-averaged changes in amplitude of the phase-locked response for the N20 m, P35 m, P60 m, and N140 m components of both Attend to MNS and Ignore MNS conditions. Overall, selective attention to the somatosensory event significantly increased the amplitude of the SI response. (B) Attention effects on latency changes in contralateral SI. Histogram representing grand-averaged changes in latency of the phase-locked response for the N20 m, P35 m, P60 m, and N140 m components of both Attend to MNS and Ignore MNS conditions. Overall, selective attention to the somatosensory event significantly reduced the latency of the SI response.
the SI virtual sensor response [F(1,11) = 38.99, p b 0.0001, ES = .78] (Fig. 2A). Overall, attention to the MNS also produced an significantly quicker virtual sensor response [F(1,11) = 9.84, p b 0.009, ES = .47] (Fig. 2B). SI-contralateral effects of attention as a function of frequency and phase In order to examine effects of somatosensory attention on the SI frequency domain, virtual sensors of selected bandwidths were created and transformed with a Morlet wavelet to produce time– frequency responses. Fig. 3A depicts the grand-averaged, phaselocked + non-phase-locked response as well as phase-locked-only and non-phase-locked-only responses. Overall, the phase-locked + non-phase-locked analysis revealed three distinct components of rhythmic changes in SI-C during both attentional conditions in response to a stimulus event: (1) an early broad-spectrum increase in power that was temporally coincident with the N20-P35-N60 phase-locked response; (2) a prolonged increase in lower band power (b13 Hz) throughout the trial; and (3) a suppression of power (eventrelated desynchrony: ERD) and subsequent rebound (event-related synchrony: ERS) in the beta band in the mid and later stages of the trial, respectively. To examine changes in power and phase with selective attention we performed statistical comparisons separately for phase-locked and non-phase-locked activities. Significant effects of attention on rhythmcity are shown in Fig. 3B. Attention to the MNS significantly increased phase-locked activity in the gamma band (30–55 Hz) centering around 20–50 ms, compared to when attention was directed away from the MNS [t(1,11) = 2.425, p b 0.05]. Moreover, selective attention to the MNS also significantly increased subsequent beta ERD and ERS activities that were non-phase-locked to stimulus onset [t(1,11) = 3.27, p b 0.01 and t(1,11) = 2.54, p b 0.05, respectively]. A moderate increase in non-phase-locked gamma activities did not reach significance [t(1,11) = 1.7, p N 0.05].
SI-ipsilateral activity Low amplitude ipsilateral SI (SI-I) activity was observed in many participants' SAM images but virtual sensor data revealed only a few detectable peaks of low amplitude (data not shown). Therefore, SI-I activity was not analyzed further. SII SII-contralateral effects of attention as a function of time Differential SAM analyses of broad-spectrum activity (1–200 Hz) in the Attend to MNS condition revealed bilateral activation of the association somatosensory cortex (SII) in the parietal operculum with stronger activation in the contralateral hemisphere (SII-C) than the ipsilateral hemisphere (SII-I) (Fig. 4A). Based on these locations, we created virtual sensors for each individual's SII-C and SII-I response, for both attentional conditions. The virtual sensors shown in Fig. 4B depict the grand-averaged power changes over time of the broadspectrum SII-C response in both attentional conditions. The stimulusrelated averaged signals were generally characterized by three components: N50 m, P100 m, and N200 m. Repeated measures ANOVA revealed a main effect of attention on SII amplitude in which attention to the MNS increased the overall power of the SII virtual sensor [F(1,11) = 22.454, p b 0.001, ES = .67] (Fig. 5A). There was no main effect of attentional condition on the latency of the SII response [F(1,11) = 3.47, p N 0.05, ES = .24] (Fig. 5B). SII-contralateral effects of attention on frequency and phase Fig. 6A depicts the grand-averaged, phase-locked + non-phaselocked response, phase-locked-only, and non-phase-locked-only responses. Overall, the phase-locked + non-phase-locked analysis showed that SII-C activity was characterized by strong, prolonged oscillations in the alpha band regardless of attentional condition. Statistical comparisons of the changes in SII-C frequency source power
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
1781
Fig. 3. (A) Attentional effects on contralateral SI oscillations. Grand-averaged contralateral SI TFRs for combined phase-locked + non-phase-locked, phase-locked-only, and nonphase-locked-only oscillations of control subjects in both attentional conditions. All plots were baselined using the average spectral energy observed in the corresponding prestimulus period (−200 to 0 ms) and represent the grand mean TFR of the individual, virtual channel, spatially-filtered single trials. Selected time–frequency boundaries are outlined on the TFR plots. Source power: A − m2 = (−n × 10−17 to +n × 10−17). (B) Statistical comparison of attentional conditions for phase-locked and non-phase-locked contralateral SI responses.
and phase between the two attentional conditions are shown in Fig. 6B. Attention to the MNS significantly increased synchrony of an early phase-locked alpha response (0–100 ms) [t(1,11) = 5.62, p b 0.001] and a subsequent non-phase-locked alpha response (100–400 ms)
[t(1,11) = 2.37, p b 0.05] compared to when the participants attended away from the somatosensory event. Some modest increases in phaselocked gamma [t(1,11) = 0.836, p N 0.05] and beta [t(1,11) = 1.12, p N 0.05] powers did not reach significance.
Fig. 4. (A) Localization of the bilateral SII response. Example of the bilateral locations of maximal SII activity during the active period following median nerve stimulation in the Attend to MNS condition. Subsequent virtual sensors and TFRs analyses were based upon these target locations. Inset shows averaged phase-locked fields from all MEG channels for control and active periods. (B) Grand-averaged virtual sensors of the contralateral SII response. Virtual sensors for changes in source power over the duration of the averaged trials were averaged across subjects.
1782
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
Fig. 5. (A) Attention effects on amplitude changes in contralateral SII. Histogram representing grand-averaged changes in amplitude of the phase-locked response for the N50 m, P100 m, and N200 m components of both Attend to MNS and Ignore MNS conditions. Overall, selective attention to the somatosensory event significantly increased the amplitude of the SII response. (B) Attention effects on latency changes in contralateral SII. Histogram representing grand-averaged changes in latency of the phase-locked response for the N50 m, P100 m, and N200 m components of both Attend to MNS and Ignore MNS conditions. Overall, selective attention to the somatosensory event did not significantly affect the latency of the SII response.
SII-ipsilateral effects of attention as a function of time Virtual sensors of the SII-I data were similar to that of SII-C although overall lower in amplitude. The first latency peak of the SII-I phase-locked response occurred at 56.2 ms ± 3.68 SEM with a peak value of 0.7 fT ± 0.43 SEM and 54.1 ms ± 7.11 SEM with a peak value of 0.79 fT ± 0.51 SEM for Attend to MNS and Ignore MNS conditions, respectively. There were no statistical differences of response time or peak values between conditions. SII-ipsilateral effects of attention on frequency and phase SII-I peaks were also characterized by strong, prolonged oscillations in the sigma band (7–9 Hz) that occurred in both attention conditions. No changes in time–frequency responses were visible in SII-I and TFR statistical analyses did not detect any differences in phase or power between conditions. Fig. 7 depicts a summary of the sequence of somatosensory oscillatory activity in SI and SII to somatosensory events and how it was influenced by attention over time. Topographical specificity of the SI and SII effects The primary visual cortex (VI) shows modulations in the alpha bandwidth in response to visual attention (Rihs et al., 2009; Fujisawa et al., 2008). Because, in our paradigms, both phase-locked and nonphase-locked SII alpha rhythms (7–9 Hz) modulated with attention we investigated participants' VI time–frequency responses to verify topographical specificity of the observed attentional effects. VI virtual sensors deriving from each participant's activation peak in the primary visual cortex were grand-averaged and transposed into TFRs. VI TFRs showed only non-phase-locked responses, consistent with the absence of repetitive, time-locked visual events. VI exhibited stronger alpha oscillations (10–12 Hz) when attention was directed to the tactile event while lower frequency oscillations (3–6 Hz) increased when participants attended to the video (and ignored the
MNS). No beta or gamma oscillations were present in either condition (see Supplementary Fig. 1). As alpha oscillations in SII and VI differed both in bandwidth and in temporal onset and offset, we propose that the somatosensory effects were highly localized and not confounded by oscillation in the visual cortex. Discussion We examined the changes in amplitude, frequency, and phase of somatosensory oscillations while manipulating selective attention. Our major findings indicate that (1) attention modulates SI phaselocked gamma rhythms and non-phase-locked beta rhythms, (2) attention modulates SII phase-locked and non-phase-locked alpha rhythms, and (3) early increases in phase-locked SI gamma precede all SII responses. Selective attention modulates power and phase of SI oscillations Gamma-band synchrony has been associated with cognitive functions such as attention, working memory, and top-down modulation of sensory processing (Fries et al., 2001; Pesaran et al., 2002). Human EEG research shows that the strength of the early phase-locked gamma response centered around 40 Hz is particularly sensitive to modulation by controlled attention (Debener et al., 2002; Fries et al., 2001; Struber et al., 2000). Our data show that selective attention to a somatosensory stimulus increased the early phaselocked SI gamma response. This increase commenced at 20 ms and was most robust around 40 Hz. Thus, the median nerve stimulation paradigm is an effective method for determining the earliest stages of attentional influence. We also compared changes in SI alpha and beta band synchrony across the two attentional conditions. In our study, SI alpha rhythms were not modulated by attention. Moreover, only alpha synchrony,
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
1783
Fig. 6. (A) Attentional effects on contralateral SII oscillations. Grand-averaged, contralateral SII time–frequency responses for combined phase-locked + non-phase-locked, phaselocked-only, and non-phase-locked-only oscillations of control subjects in both attentional conditions. Due to very low-frequency noise (2–4 Hz, throughout duration of trial), all plots were baselined using the average spectral energy observed across the trial (−670 to 0 ms). Each plot represents the grand mean TFR of the individual, virtual channel, spatiallyfiltered single trials. Selected time–frequency boundaries are outlined on the TFR plots. Source power: A − m2 = (−n × 10−17 to +n × 10−17). (B) Statistical comparison of attentional conditions for phase-locked and non-phase-locked contralateral SII responses.
but not desynchrony, was present. A large number of studies within the visual and auditory domains propose that alpha ERD reflects sensory processing while alpha ERS reflects cortical inhibition
(Sauseng et al., 2005; Worden et al., 2000; Thut et al., 2006; Rihs et al., 2007). It may be that somatosensory alpha activities may not show the same consistent relationship as the visual and auditory
Fig. 7. Summary of how attention affects phase-locked and non-phase-locked somatosensory oscillations. Electrical stimulation of the right median nerve elicits a robust, contralateral response in the left cortical hemisphere. Neuromagnetic recordings and subsequent beamformer analysis show that the two peak activities localize in contralateral SI and SII. Waveform analysis combined with time–frequency transformations indicate that the initial SI response shows amplitude peaks at 20, 35, and 60 ms following stimulus onset which coincide with phase-locked gamma activity and increase with attention to the somatosensory event. SII amplitude peaks that occur at 55 and 100 ms post-stimulus coincide with phase-locked alpha activity. All phase-locked activities are amplified by attention. Both SI and SII also show non-phase-locked oscillations in response to the somatosensory event that begin ∼100 ms after stimulus onset (beta and alpha rhythms, respectively). All non-phase-locked components are amplified with selective attention.
1784
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
systems do. Alternatively, the robust alpha synchrony we observed may be specific to the electrical stimulus applied. Correspondingly, Forss et al. (2001), Palva et al. (2005), and Dockstader et al. (2009) showed that electrical stimulation that was applied to the wrist (Forss et al.; Dockstader et al.) or the index finger (Palva et al.) elicited an immediate and robust increase in SI alpha power. This was later followed by an alpha ERD when stimulus frequency was 0.5 Hz or slower (Palva et al.; Dockstader et al.) but not when stimuli were presented rapidly (1.5–12 Hz). In our experiment (and in the case of Forss et al.) the absence of a later alpha ERD response may simply be due to the brief time period between stimulus events. Although attention did not alter SI alpha synchrony, both nonphase-locked beta ERD and ERS occurred. Sensorimotor beta oscillations have largely been associated with motor activities (Neuper and Pfurtscheller, 1996; Sochurkova et al., 2006). However, these oscillations may play a role in attention as increases in sensorimotor beta power correlate with attention directed toward a motoric event (Muthukumaraswamy and Singh, 2008) and improved sensorimotor performance (Egner and Gruzelier, 2004; Egner et al., 2004; Vernon et al., 2003). Interestingly, individuals with attention-deficit/hyperactivity disorder show diminished sensorimotor beta response to somatosensory input (Dockstader et al., 2008). Selective attention modulates power and phase of alpha (‘sigma') oscillations in SII Here we show that attention to a somatosensory event does modulate SII alpha ERS. Specifically, there was an immediate, stimulus-locked increase in 7–9 Hz synchrony when participants attended to the somatosensory event followed by a delayed and prolonged non-phase-locked increase in neural synchrony. To our knowledge, this is the first report of attention-related changes in alpha oscillations observed in human SII. This results suggests that, in SII, there is a direct and active role for alpha ERS in somatosensory perception and attention. Moreover, while the dynamics of the alpha rhythm remain unclear, this result contributes to the recent theory that cortical inhibition is only one of several roles of alpha ERS (Palva and Palva, 2007). Because SII alpha activity was restricted to a 7–9 Hz bandwidth, we suggest that the observed SII alpha activity may be related to the previously identified ‘sigma’ rhythm—an alpha rhythm variant restricted to 7–9 Hz in SII originally reported by Narici et al. (2001) and Forss et al. (2001).
We suggest that our observed sequence of phase-locked events reflects neural communication of sensory information from SI to SII that is enhanced, in both regions, with attention to the somatosensory event. Selective attention to the somatosensory stimulus also strengthened non-phase-locked SI beta ERD and ERS and SII alpha, or sigma, ERS. As the onset of non-phase-locked events began after SI and SII phase-locked activities were completed it is unlikely that the non-phase-locked oscillations reflected somatosensory stimulus perception per se. We suggest, rather, that these rhythms reflect cognitive processes driven by stimulus input and modulated by attention. Hemispheric asymmetry Attentional effects were found to be lateralized for both SI and SII, suggesting that the contralateral, rather than ipsilateral, hemisphere is involved in selective attention to somatosensory stimuli. Cortical hemispheric asymmetry has been associated with attention in healthy participants (Floel et al., 2005a; Floel et al., 2005b; Floel et al., 2001). It is also possible that lateralization effects might differ if we conducted the task again with left-handers or in right-handers while stimulating their left median nerve. However, an fMRI study showed a functional dominance of the left SII in secondary somatosensory cortex, regardless of handedness, using median nerve stimulation (Simoes and Hari, 1999). Hemispheric asymmetry may also be dependent on the particular stimulus applied as recent fMRI studies have shown that ipsilateral somatosensory responses are stronger for mechanical versus median nerve electrical stimulation (Lipton et al., 2006). Future directions Although our study revealed several novel findings, there are several remaining questions. To answer these questions, future studies may include (1) coherence analyses in which the relationship between SI gamma and SII alpha may be further defined, (2) investigating neural somatosensory responses to natural stimuli, (3) examining the relationship between enhanced neural somatosensory response and enhanced task performance, (4) determining whether early attentional changes to somatosensory input in SI and SII are independent of the priming sensory modality, and (5) applying our paradigm to individuals with attention deficits. Conclusion
Dissociating phase-locked and non-phase-locked oscillatory changes Phase-locked and non-phase-locked changes in synchrony can reflect functionally separate cognitive processes. Whereas it is suggested that phase-locked activities reflect enhanced information transfer between brain areas in response to an external event, nonphase-locked activities more likely reflect top-down influences that modify the sensitivity of a region to incoming events and (Klimesch et al., 2007; Pineda, 2005). Changes in either of these forms of activity can enhance task-related performance. For example, increases in alpha phase-locking in the visual cortex correlate with more sensitive perception (Hanslmayr et al., 2005) whereas changes in somatosensory alpha non-phase-locked activities correlate with learning (Zhuang et al., 1997) and anticipation (Babiloni et al., 2004). However, these rhythms do not appear to be critical in the perception of a somatosensory event (Dockstader et al., 2009). In our study, selective attention to a somatosensory stimulus enhanced synchrony in SI phase-locked gamma (∼20 ms) and SII phase-locked alpha ∼30 ms following the SI response. Postsynaptic targets are particularly sensitive to gamma band synchrony (Womelsdorf et al., 2006) and therefore the increase in SI gamma synchrony should indeed be followed by an increase in synchrony of the downstream SII neurons.
This is the first study to uncouple phase-locked versus non-phaselocked oscillations in SI during selective attention. It is also the first study to examine time–frequency characteristics of downstream SII activities. We suggest that early phase-locked oscillations signal the relay of perceptual input from SI to SII. Later non-phase-locked oscillations are sensitive to selective attention and may signal heightened stimulus processing. Understanding the dynamics of cortical oscillations and their relationship to attentional control will provide new insights into some of the earliest and most basic aspects of cognitive processing. Acknowledgments We thank Dr. Margot Taylor for feedback on the manuscript and Travis Mills, Sonya Bells, and Christine Popovich for their technical support. This research was supported in part by funds from a Canadian Institutes of Health Research Operating Grant (CIHR # 64279; PI is D.C.) and an operating grant from The Hospital for Sick Children Psychiatry Endowment Fun (C.D., D.C., and R.T.), as well as an Ontario Mental Health Foundation fellowship (C.D.) and the Canada Research Chairs Program (R.T.).
C. Dockstader et al. / NeuroImage 49 (2010) 1777–1785
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.neuroimage.2009.09.035. References Babiloni, C., Brancucci, A., Arendt-Nielsen, L., Babiloni, F., Capotosto, P., Carducci, F., Cincotti, F., Romano, L., Chen, A.C., Rossini, P.M., 2004. Alpha event-related desynchronization preceding a go/no-go task: a high-resolution EEG study. Neuropsychology 18, 719–728. Bauer, M., Oostenveld, R., Peeters, M., Fries, P., 2006. Tactile spatial attention enhances gamma-band activity in somatosensory cortex and reduces low-frequency activity in parieto-occipital areas. J. Neurosci. 26, 490–501. Bekisz, M., Wrobel, A., 2003. Attention-dependent coupling between beta activities recorded in the cat's thalamic and cortical representations of the central visual field. Eur. J. Neurosci. 17, 421–426. Broring, T., Rommelse, N., Sergeant, J., Scherder, E., 2008. Sex differences in tactile defensiveness in children with ADHD and their siblings. Dev. Med. Child Neurol. 50, 129–133. Cheyne, D., Bostan, A.C., Gaetz, W., Pang, E.W., 2007. Event-related beamforming: a robust method for presurgical functional mapping using MEG. Clin. Neurophysiol. 118, 1691–1704. Debener, S., Kranczioch, C., Herrmann, C.S., Engel, A.K., 2002. Auditory novelty oddball allows reliable distinction of top-down and bottom-up processes of attention. Int. J. Psychophysiol. 46, 77–84. Dockstader, C., Gaetz, W., Cheyne, D., Wang, F., Castellanos, F.X., Tannock, R., 2008. MEG event-related desynchronization and synchronization deficits during basic somatosensory processing in individuals with ADHD. Behav. Brain Funct. 4, 8. Dockstader, C., Gaetz, W., Cheyne, D., Tannock, R., 2009. Abnormal neural reactivity to unpredictable sensory events in attention-deficit/hyperactivity disorder. Biol. Psychiatry 66, 376–383. Driver, J., Frackowiak, R.S., 2001. Neurobiological measures of human selective attention. Neuropsychologia 39, 1257–1262. Egner, T., Gruzelier, J.H., 2004. EEG biofeedback of low beta band components: frequency-specific effects on variables of attention and event-related brain potentials. Clin. Neurophysiol. 115, 131–139. Egner, T., Zech, T.F., Gruzelier, J.H., 2004. The effects of neurofeedback training on the spectral topography of the electroencephalogram. Clin. Neurophysiol. 115, 2452–2460. Engel, A.K., Fries, P., Singer, W., 2001. Dynamic predictions: oscillations and synchrony in top-down processing. Nat. Rev. Neurosci. 2, 704–716. Fahle, M., 2009. Perceptual learning and sensomotor flexibility: cortical plasticity under attentional control? Philos. Trans. R. Soc. Lond., B Biol. Sci. 364, 313–319. Fell, J., Fernandez, G., Klaver, P., Elger, C.E., Fries, P., 2003. Is synchronized neuronal gamma activity relevant for selective attention? Brain Res. Brain Res. Rev. 42, 265–272. Floel, A., Knecht, S., Lohmann, H., Deppe, M., Sommer, J., Drager, B., Ringelstein, E.B., Henningsen, H., 2001. Language and spatial attention can lateralize to the same hemisphere in healthy humans. Neurology 57, 1018–1024. Floel, A., Buyx, A., Breitenstein, C., Lohmann, H., Knecht, S., 2005a. Hemispheric lateralization of spatial attention in right- and left-hemispheric language dominance. Behav. Brain Res. 158, 269–275. Floel, A., Jansen, A., Deppe, M., Kanowski, M., Konrad, C., Sommer, J., Knecht, S., 2005b. Atypical hemispheric dominance for attention: functional MRI topography. J. Cereb. Blood. Flow Metab. 25, 1197–1208. Forss, N., Narici, L., Hari, R., 2001. Sustained activation of the human SII cortices by stimulus trains. NeuroImage 13, 497–501. Fries, P., Reynolds, J.H., Rorie, A.E., Desimone, R., 2001. Modulation of oscillatory neuronal synchronization by selective visual attention. Science 291, 1560–1563. Fujisawa, J., Touyama, H., Hirose, M., 2008. Extracting alpha band modulation during visual spatial attention without flickering stimuli using common spatial pattern. Conf. Proc. IEEE Eng. Med. Biol. Soc. 620–623. Georgiou-Karistianis, N., Howells, D., Bradshaw, J., 2003. Orienting attention in obsessive-compulsive disorder. Cogn. Behav. Neurol. 16, 68–74. Hanslmayr, S., Klimesch, W., Sauseng, P., Gruber, W., Doppelmayr, M., Freunberger, R., Pecherstorfer, T., 2005. Visual discrimination performance is related to decreased alpha amplitude but increased phase locking. Neurosci. Lett. 375, 64–68. Herrmann, C.S., Knight, R.T., 2001. Mechanisms of human attention: event-related potentials and oscillations. Neurosci. Biobehav. Rev. 25, 465–476. Hillyard, S.A., 1993. Electrical and magnetic brain recordings: contributions to cognitive neuroscience. Curr. Opin. Neurobiol. 3, 217–224. Klimesch, W., Sauseng, P., Hanslmayr, S., 2007. EEG alpha oscillations: the inhibitiontiming hypothesis. Brain Res. Rev. 53, 63–88. Knudsen, E.I., 2007. Fundamental components of attention. Annu. Rev. Neurosci. 30, 57–78. Kok, A., Ridderinkhof, K.R., Ullsperger, M., 2006. The control of attention and actions: current research and future developments. Brain Res. 1105, 1–6. Lipton, M.L., Fu, K.M., Branch, C.A., Schroeder, C.E., 2006. Ipsilateral hand input to area 3b revealed by converging hemodynamic and electrophysiological analyses in macaque monkeys. J. Neurosci. 26, 180–185. Muthukumaraswamy, S.D., Singh, K.D., 2008. Modulation of the human mirror neuron system during cognitive activity. Psychophysiology 45, 896–905.
1785
Naatanen, R., Michie, P.T., 1979. Early selective-attention effects on the evoked potential: a critical review and reinterpretation. Biol. Psychol. 8, 81–136. Narici, L., Forss, N., Jousmaki, V., Peresson, M., Hari, R., 2001. Evidence for a 7- to 9-Hz “sigma” rhythm in the human SII cortex. NeuroImage 13, 662–668. Neuper, C., Pfurtscheller, G., 1996. Post-movement synchronization of beta rhythms in the EEG over the cortical foot area in man. Neurosci. Lett. 216, 17–20. Neuper, C., Wortz, M., Pfurtscheller, G., 2006. ERD/ERS patterns reflecting sensorimotor activation and deactivation. Prog. Brain Res. 159, 211–222. Niebur, E., 2002. Electrophysiological correlates of synchronous neural activity and attention: a short review. Biosystems 67, 157–166. Niebur, E., Hsiao, S.S., Johnson, K.O., 2002. Synchrony: a neuronal mechanism for attentional selection? Curr. Opin. Neurobiol. 12, 190–194. Palva, S., Palva, J.M., 2007. New vistas for alpha-frequency band oscillations. Trends Neurosci. 30, 150–158. Palva, S., Linkenkaer-Hansen, K., Naatanen, R., Palva, J.M., 2005. Early neural correlates of conscious somatosensory perception. J. Neurosci. 25, 5248–5258. Parush, S., Sohmer, H., Steinberg, A., Kaitz, M., 1997. Somatosensory functioning in children with attention deficit hyperactivity disorder. Dev. Med. Child Neurol. 39, 464–468. Parush, S., Sohmer, H., Steinberg, A., Kaitz, M., 2007. Somatosensory function in boys with ADHD and tactile defensiveness. Physiol. Behav. 90, 553–558. Pesaran, B., Pezaris, J.S., Sahani, M., Mitra, P.P., Andersen, R.A., 2002. Temporal structure in neuronal activity during working memory in macaque parietal cortex. Nat. Neurosci. 5, 805–811. Pineda, J.A., 2005. The functional significance of mu rhythms: translating “seeing” and “hearing” into “doing”. Brain Res. Brain Res. Rev. 50, 57–68. Ray, S., Niebur, E., Hsiao, S.S., Sinai, A., Crone, N.E., 2008. High-frequency gamma activity (80–150 Hz) is increased in human cortex during selective attention. Clin. Neurophysiol. 119, 116–133. Rihs, T.A., Michel, C.M., Thut, G., 2007. Mechanisms of selective inhibition in visual spatial attention are indexed by alpha-band EEG synchronization. Eur. J. Neurosci. 25, 603–610. Rihs, T.A., Michel, C.M., Thut, G., 2009. A bias for posterior alpha-band power suppression versus enhancement during shifting versus maintenance of spatial attention. Neuroimage 44 (1), 190–199. Robinson, S., Vrba, J., 1999. Functional neuroimaging by synthetic aperture magnetometry (SAM). Tohku Univ. Press, Sendai, Japan. Sauseng, P., Klimesch, W., Stadler, W., Schabus, M., Doppelmayr, M., Hanslmayr, S., Gruber, W.R., Birbaumer, N., 2005. A shift of visual spatial attention is selectively associated with human EEG alpha activity. Eur. J. Neurosci. 22, 2917–2926. Scherder, E.J., Rommelse, N.N., Broring, T., Faraone, S.V., Sergeant, J.A., 2008. Somatosensory functioning and experienced pain in ADHD-families: a pilot study. Eur. J. Paediatr. Neurol. 12 (6), 461–469. Sekihara, K., Nagarajan, S.S., Poeppel, D., Marantz, A., 2004. Asymptotic SNR of scalar and vector minimum-variance beamformers for neuromagnetic source reconstruction. IEEE Trans. Biomed. Eng. 51, 1726–1734. Simoes, C., Hari, R., 1999. Relationship between responses to contra- and ipsilateral stimuli in the human second somatosensory cortex SII. NeuroImage 10, 408–416. Sochurkova, D., Rektor, I., Jurak, P., Stancak, A., 2006. Intracerebral recording of cortical activity related to self-paced voluntary movements: a Bereitschaftspotential and event-related desynchronization/synchronization. SEEG study. Exp. Brain Res. 173, 637–649. Stancak, A., 2006. Cortical oscillatory changes occurring during somatosensory and thermal stimulation. Prog. Brain Res. 159, 237–252. Struber, D., Basar-Eroglu, C., Hoff, E., Stadler, M., 2000. Reversal-rate dependent differences in the EEG gamma-band during multistable visual perception. Int. J. Psychophysiol. 38, 243–252. Tallon-Baudry, C., Bertrand, O., Delpuech, C., Pernier, J., 1996. Stimulus specificity of phase-locked and non-phase-locked 40 Hz visual responses in human. J. Neurosci. 16, 4240–4249. Thut, G., Nietzel, A., Brandt, S.A., Pascual-Leone, A., 2006. Alpha-band electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts visual target detection. J. Neurosci. 26, 9494–9502. Varela, F., Lachaux, J.P., Rodriguez, E., Martinerie, J., 2001. The brainweb: phase synchronization and large-scale integration. Nat. Rev. Neurosci. 2, 229–239. Vernon, D., Egner, T., Cooper, N., Compton, T., Neilands, C., Sheri, A., Gruzelier, J., 2003. The effect of training distinct neurofeedback protocols on aspects of cognitive performance. Int. J. Psychophysiol. 47, 75–85. Womelsdorf, T., Fries, P., Mitra, P.P., Desimone, R., 2006. Gamma-band synchronization in visual cortex predicts speed of change detection. Nature 439, 733–736. Worden, M.S., Foxe, J.J., Wang, N., Simpson, G.V., 2000. Anticipatory biasing of visuospatial attention indexed by retinotopically specific alpha-band electroencephalography increases over occipital cortex. J. Neurosci. 20, RC63. Wrobel, A., 2000. Beta activity: a carrier for visual attention. Acta Neurobiol. Exp. (Wars) 60, 247–260. Wrobel, A., Ghazaryan, A., Bekisz, M., Bogdan, W., Kaminski, J., 2007. Two streams of attention-dependent beta activity in the striate recipient zone of cat's lateral posterior–pulvinar complex. J. Neurosci. 27, 2230–2240. Zhuang, P., Toro, C., Grafman, J., Manganotti, P., Leocani, L., Hallett, M., 1997. Eventrelated desynchronization (ERD) in the alpha frequency during development of implicit and explicit learning. Electroencephalogr. Clin. Neurophysiol. 102, 374–381.