Modulation of Alpha Activity in the Parieto-occipital Area by Distractors during a Visuospatial Working Memory Task: A Magnetoencephalographic Study Satoe Ichihara-Takeda1, Shogo Yazawa1, Takashi Murahara1, Takanobu Toyoshima1, Jun Shinozaki1, Masanori Ishiguro1, Hideaki Shiraishi2, Nozomu Ikeda1, Kiyoji Matsuyama1, Shintaro Funahashi3, and Takashi Nagamine1
Abstract ■ Oscillatory brain activity is known to play an essential role in
information processing in working memory. Recent studies have indicated that alpha activity (8–13 Hz) in the parieto-occipital area is strongly modulated in working memory tasks. However, the function of alpha activity in working memory is open to several interpretations, such that alpha activity may be a direct neural correlate of information processing in working memory or may reflect disengagement from information processing in other brain areas. To examine the functional contribution of alpha activity to visuospatial working memory, we introduced visuospatial distractors during a delay period and examined neural activity from the whole brain using magnetoencephalography.
INTRODUCTION Working memory is a mechanism for the short-term active storage of information as well as for the processing of stored information (Baddeley, 1986). Goldman-Rakic (1987) proposed that working memory is an important concept for understanding dorsolateral prefrontal cortex (DLPFC) functions. Single-cell recordings and functional MRI studies have demonstrated that excitatory sustained activity is observed in the DLPFC during the delay period in the visuospatial working memory task (Ichihara-Takeda & Funahashi, 2007; Roesch & Olson, 2005; Curtis & DʼEsposito, 2003; Sakai, Rowe, & Passingham, 2002; Funahashi, Bruce, & Goldman-Rakic, 1989). Therefore, this sustained activity has been considered to be a neural correlate of the temporary maintenance of information. However, this sustained activity is observed not only in the DLPFC but also in several other brain areas, including the posterior parietal cortex (Sakai et al., 2002; Chafee & Goldman-Rakic, 1998), occipital visual association areas (Carlson et al., 1998), the caudate nucleus (Postle & DʼEsposito, 1999), and the
1
Sapporo Medical University, Japan, 2Hokkaido University School of Medicine, Sapporo, Japan, 3Kyoto University © 2015 Massachusetts Institute of Technology
The strength of event-related alpha activity was estimated using the temporal spectral evolution (TSE) method. The results were as follows: (1) an increase of alpha activity during the delay period as indicated by elevated TSE curves was observed in parietooccipital sensors in both the working memory task and a control task that did not require working memory; and (2) an increase of alpha activity during the delay period was not observed when distractors were presented, although TSE curves were constructed only from correct trials. These results indicate that the increase of alpha activity is not directly related to information processing in working memory but rather reflects the disengagement of attention from the visuospatial input. ■
thalamic mediodorsal nucleus ( Watanabe & Funahashi, 2004). These results indicate that several brain areas participate in working memory. Therefore, to better understand the neural mechanisms of working memory, it is important to measure neuronal activity from the whole brain with high temporal resolution while participants perform working memory tasks. Electroencephalography and magnetoencephalography (MEG) can be used to measure whole-brain neuronal activity with high temporal resolution. Studies that have used these techniques suggest that oscillatory alpha activity in posterior cortical areas reflects an idling condition of the cortex (Pfurtscheller, Stancák, & Neuper, 1996) or a preparatory process for attention (Thut, Nietzel, Brandt, & Pascual-Leone, 2006; Worden, Foxe, Wang, & Simpson, 2000). Oscillatory alpha activity has also been analyzed to determine the neuronal processes that are necessary for working memory. Alpha activity is strongly modulated during working memory tasks, and this activity is often observed in the parieto-occipital area in addition to the prefrontal area (Ciesielski et al., 2007; Jokisch & Jensen, 2007; Tuladhar et al., 2007; Jensen, Gelfand, Kounios, & Lisman, 2002). Especially, the increase of alpha activity in the parieto-occipital area is thought to reflect either an Journal of Cognitive Neuroscience 27:3, pp. 453–463 doi:10.1162/jocn_a_00718
active neural process related to memory maintenance (Palva & Palva, 2007; Palva, Palva, & Kaila, 2005) or the active inhibition of posterior cortical regions that are not required for the task (Jensen & Mazaheri, 2010; Klimesch, Sauseng, & Hanslmayr, 2007). Recent studies have suggested that parieto-occipital alpha activity during the delay period of working memory tasks reflects the disengagement or inhibition of the visual dorsal stream (Jokisch & Jensen, 2007; Tuladhar et al., 2007). Individual differences of alpha activity appear to correlate with individual differences of the capacity for suppressing irrelevant information (Sauseng et al., 2009). However, the contribution of parieto-occipital alpha activity to visuospatial working memory processes continues to be debated. Therefore, in this study, we examined the signals or factors that affect the strength of alpha activity observed during the delay period of a visuospatial working memory task using MEG. We analyzed the temporal pattern of the strength of alpha activity in the whole brain in both working memory tasks and control tasks that did not require working memory. In addition, we introduced visuospatial distractors during the delay period of the task to estimate the effect of alpha activity on the maintenance of information in working memory.
METHODS Participants The experiments were carried out in 10 healthy participants (seven men, nine right-handed and one left-handed, age range = 21–49 years, mean = 30.9 years). Informed consent was obtained from each participant before the study. The study protocol was approved by the ethics committees of Sapporo Medical University and Hokkaido University. Task In this study, we used two tasks (a delayed matching-tosample task [DMST] and a control task [CONT]) under two distractor conditions (without and with distractors). All participants performed the tasks under four conditions (each task without and with distractors: DMST and DMST +d, CONT and CONT+d). Trials under these four conditions consisted of five periods: fixation period (3 sec), cue period (0.5 sec), delay period (5 sec), end-signal period (0.5 sec), and probe period (0.8 sec). Visual stimuli were presented on a back-projection screen placed 1.4 m from the participant by an LCD projector (LT260SK; NEC, Tokyo, Japan). Each trial started with the presentation of Japanese characters (hiragana; white, 1.5° in visual angle). The characters were presented at the center of the screen to enforce the participantsʼ maintenance of eye fixation at the center of the screen and to prevent the use of language to maintain spatial information of the visual cue during the delay 454
Journal of Cognitive Neuroscience
period (Park, Holzman, & Goldman-Rakic, 1995; Park & Holzman, 1992). Participants were required to view the Japanese characters and read them silently throughout the fixation, cue, and delay periods (Figure 1). The Japanese characters were randomly chosen from a pool of 50 characters and presented successively, one by one, for 1 sec each. After the 3-sec fixation period, a visual cue (white circle, 2.5° in visual angle) was presented for 0.5 sec at a peripheral location simultaneously with the presentation of a Japanese character at the center (cue period). The location of the visual cue was randomly selected from eight predetermined positions. The 5-sec delay period then began. At the end of the delay period, the central Japanese character was replaced with a star figure, which signaled the end of the delay period, and remained on the screen for 0.5 sec (end-signal period). After the end-signal period, a probe cue was presented for 0.8 sec (probe period). In the DMST condition, participants were required to make a same–different judgment based on whether the visual cue and the probe cue appeared at the same position by lifting either the right index finger (same position) or the left index finger (different position; Figure 1A). The visual cue and the probe cue appeared at the same location in 12.5% of trials. In the CONT condition, participants were required to lift the left index finger when the probe cue was presented, regardless of its location. Therefore, they did not need to remember the position of the visual cue. After the participants responded, an intertrial interval of 3 sec was introduced. During the intertrial interval, participants were allowed to blink. In the DMST with distractor condition (DMST+d), distractors were presented during the delay period. Each distractor (white circle, 2.5° in visual angle) was presented for 0.5 sec at a location that was randomly selected from among the eight predetermined positions. The first distractor was presented 1 sec after the start of the delay period, and three other distractors were presented sequentially every 1 sec during the delay period. Their color and shape were the same as those of the visual cue (Figure 1B). In the CONT with distractor condition (CONT+d), participants did not need to remember the position of the visual cue, and distractors were presented according to the same procedure as in the DMST+d condition. The participants were exposed to four conditions (DMST, CONT, DMST+d, and CONT+d) on 2 separate days, and the order was randomized among participants. On the first day, participants performed two different conditions twice, for example, DMST (first session), DMST+d (first session), DMST (second session), and DMST+d (second session). Each session included 60 trials. On the second day, the participants performed the remaining two conditions twice, for example, CONT+d (first session), CONT (first session), CONT+d (second session), and CONT (second session). Volume 27, Number 3
Figure 1. The experimental paradigm. Two variants of a modified DMST were used: DMST with no distractor (DMST, A) and with distractors (DMST+d, B). (A) Each trial began with the presentation of Japanese characters. From the start of the fixation period until the end signal, the Japanese characters (hiragana) were presented at the center of the screen to help the participant maintain eye fixation at the center of the screen and to prevent the use of language to maintain spatial information during the delay period. After the 3-sec fixation period, a visual cue was presented for 0.5 sec (cue period). The location of the visual cue was randomly selected from eight predetermined positions. The 5-sec delay period (delay period) then began. At the end of the delay period, the central Japanese character was replaced with a star figure, which remained on the screen for 0.5 sec (end-signal period). After the end-signal period, a visual probe stimulus was presented for 0.8 sec (probe period). Participants were requested to make a same–different judgment by lifting their right index finger for “same” if the probe appeared at the same position as the cue. Participants were also requested to make a judgment by lifting their left index finger for “different” if the probe appeared in a different position than the cue. (B) In the DMST+d condition, distractors were presented during the delay period. Each distractor was presented for 0.5 sec at a location that was randomly selected from eight predetermined positions. The first distractor was presented 1 sec after the start of the delay period, and three other distractors were presented sequentially every 1 sec during the delay period. The color and shape of the distractors were the same as those of the visual cue. (C) Timing and display area for the presentation of Japanese characters (hiragana), cue, end signal, probe, and distractors during the task.
Ichihara-Takeda et al.
455
MEG Recordings We used MEG with whole-head 204-channel planar gradiometers (Vectorview; Elekta, Stockholm, Sweden) in a magnetically shielded room. The sampling frequency was 600 Hz, and the bandpass filter was set at 0.03–200 Hz. Eye movements and eye blinks were monitored by a vertical EOG. The participantsʼ motor performance (extension of the index finger) was continuously monitored by an online electromyogram recorded from the extensor indicis muscle. Anatomical landmark positions (nasion and preauricular points) were determined by a Polhemus Fastrak 3-D digitizer (Colchester, Vermont). Because the sessions were separated by 2- to 3-min breaks, each participantʼs head position relative to the MEG sensor array was measured at the beginning of each session. Before recording on the second day, each participantʼs head position relative to the MEG sensor array was measured, and the position of their head was adjusted so that the displacement of head position from the first recording was within 10 mm.
et al., 2005; Schnitzler, Salenius, Salmelin, Jousmäki, & Hari, 1997). The following time windows were analyzed: (1) −2 to 0 sec with respect to the onset of the cue period (baseline period), (2) 0–0.5 sec after the onset of the cue period (cue period), and (3) 0.5–5.5 sec after the onset of the cue period (delay period). For the purposes of this study, the delay period was of particular interest. The mean values of TSE levels over the baseline and delay periods were calculated for the most reactive sensor of the gradiometer. TSE% was defined as the percentage increase in band power during the delay period as compared with the baseline period: TSE% = (band power, delay period)/(band power, baseline period) × 100. This value was compared across four conditions using twoway repeated-measures ANOVA with two tasks (DMST and CONT) and two kinds of stimuli (without and with distractors) as factors.
RESULTS Data Analysis Epochs from 2 sec before the onset of the cue period to 6.8 sec after the onset of the cue period were extracted from continuous MEG raw data. The activity during each epoch was visually inspected, and trials that included large eye movements or blinking were excluded from the analysis. Trials in which participants made incorrect judgments were also excluded from the analysis. The level of event-related alpha activity as a function of time was estimated using the temporal spectral evolution (TSE) method (Nagamine, Kajola, Salmelin, Shibasaki, & Hari, 1996; Salmelin, Hämäläinen, Kajola, & Hari, 1995; Salmelin & Hari, 1994). TSE analysis was adopted to characterize the reactivity of the spontaneous cortical activity. In this analysis, spontaneous activity was extracted by using the frequency filter of interest. Thereafter, its temporal profile with respect to events of stimulus or movement onset was obtained by averaging the absolute value of filtered data across events. In our data analysis, the continuous MEG signals were filtered through pass bands of 8–13 Hz and then rectified. The absolute values within the selected epochs were averaged over the trials with respect to the onset of the cue period. Because we were mainly interested in the slow component with a temporal resolution of 100 msec, we smoothed the data with a 10-Hz low-pass filter. In this study, approximately 100–120 trials in each condition were averaged with respect to the onset of the cue period. However, only 60 trials were averaged for Participant 7 because of the frequent occurrence of artifacts during epochs used for analysis. Because MEG signals from planar gradiometers are strongest when the sensors are located above cortical current sources, the data from the most reactive sensor that showed the strongest TSE response were used to evaluate the effects of various experimental conditions (Ichikawa et al., 2007; Tamura 456
Journal of Cognitive Neuroscience
Behavioral Data To examine whether behavioral performance was affected by distractors during the delay period, the percentage of correct performance was calculated for each participant. Figure 2 shows the percentage of correct performance in the DMST and DMST+d conditions for all 10 participants. In five participants, the percentage of correct performance in the DMST+d condition was lower than that in the DMST condition. Three participants showed the same percentage in these two conditions, and two showed a slightly higher percentage in the DMST+d condition than in the DMST condition. The mean correct percentage among the 10 participants was 97.4% (SD = 2.3) for the
Figure 2. Behavioral data during MEG sessions. The percentage of correct performance in the DMST and DMST+d conditions is plotted for all 10 participants.
Volume 27, Number 3
DMST condition and 94.2% (SD = 7.3) for the DMST+d condition. Although the overall performance in the DMST condition was slightly better than that in the DMST+d condition, this difference was not significant (Wilcoxon ranksum test, p = .18). Alpha Activity in the DMST and the Control Task Figure 3A shows a representative participantʼs TSE curves for alpha-band oscillations constructed from the recordings of 204 sensors. The TSE curve that showed the largest change in alpha activity was observed for the parieto-occipital sensors during the DMST condition. Figure 3B shows TSE curves obtained from the most reactive sensor that showed the strongest alpha activity. Temporal modulation was clearly observed during the delay period in the DMST condition. The activity increased approximately 1 sec after the onset of the cue period (0.5 sec after the onset of the delay period). The activity peaked at 3.3 sec after the onset of the cue period and then gradually became weaker. At the end-signal period, the activity rapidly diminished and reached its lowest level in the first half of the probe period, which was lower than that during the baseline period. The activity then rapidly increased starting 0.4 sec after presentation of the probe and continued to increase beyond the end of the probe period. We observed similar temporal patterns of alpha activity during the delay period in the DMST condition in 7 of the 10 participants. However, in the remaining three participants, the modulation of alpha activity was not observed during any task epoch at any cortical location in any of the conditions. The percentage of correct performance during the DMST condition in the seven participants who showed modulation was 97.4% (SD = 2.3) and that in the three participants who did not show modulation was 97.3% (SD = 2.7). Thus, there was no difference in the correct performance rate between participants who did and did not show an increase of alpha activity during the delay period of the DMST condition. Figure 4 shows TSE curves obtained from the most reactive sensor that showed the strongest alpha activity in these seven participants. In all of the participants, these sensors were located in the parieto-occipital area. Alpha activity started to increase after the initiation of the delay period in the DMST condition in Participants 1 and 2. In the remaining five participants (Participants 3–7), alpha activity was first suppressed during the cue period and then increased during the delay period in the DMST condition. An increase of alpha activity similar to the increase of activity during the delay period in the DMST condition was observed in the CONT condition (Participants 2, 4, 5, 6, and 7). The alpha activity in the CONT condition as a percentage of that in the DMST condition in Participants 2, 4, 5, 6, and 7 was 95.1%, 90.7%, 98.3%, 96.6%, and 103%, respectively. However, in Participants 1 and 3,
the alpha activity in the CONT condition as a percentage of that in the DMST condition was 78.2% and 79.2%, respectively. Thus, alpha activity during the delay period increased in the CONT condition as much as in the DMST condition in these five participants. The suppression of alpha activity during the cue period was also observed in the CONT condition in all participants who showed similar suppression in the DMST condition. Furthermore, in both the DMST and CONT conditions, all seven participants demonstrated the rapid suppression of alpha activity after the onset of the endsignal period. The alpha activity reached the lowest value approximately 0.3 sec after the onset of the probe period. This activity then started increasing rapidly approximately 0.4 sec after the onset of the probe period. In addition, the transient suppression of increased alpha activity was observed from 2, 3, 4, and 5 sec after the onset of the cue stimulus in six participants in the DMST and CONT conditions. In this task, Japanese characters were presented successively, one by one, for 1 sec each. The timing of the transient suppression of alpha activity corresponded to that when a Japanese character was presented. Thus, the suppression of activity coincided with the presentation of visual stimuli (cue, end signal, probe, and Japanese characters). We should also stress that this suppression was observed not only during the visuospatial working memory task but also during the control task that did not require working memory. On the other hand, the temporal pattern of the increase of alpha activity after the suppression caused by cue presentation was different from that caused by presentation of the probe stimulus. After the suppression caused by cue presentation, it took about 1.5 sec for alpha activity to return to the baseline level. On the other hand, a rapid increase of activity was observed in the middle of the probe period. A Japanese character was presented 1 sec after the onset of the cue stimulus (0.5 sec after the onset of the delay period), whereas no other visual stimulus was presented after the onset of the probe stimulus. Thus, the timing of the slow increase of alpha activity after the suppression caused by cue presentation corresponded to that when a Japanese character was presented. In the DMST condition, five of seven participants exhibited the suppression of alpha activity during the cue period. To identify the brain regions that were responsible for this suppression of alpha activity during the cue period, the cortical distribution was compared between the suppression of alpha activity as shown by TSE curves and event-related cue period activity. Twenty sensors were selected for this comparison from among sensors adjacent to the sensor that exhibited the strongest alpha activity according to TSE curves in the DMST condition (Figure 5). Figure 5A illustrates examples of TSE curves obtained from sensors within the parieto-occipital area where the suppression of alpha activity was observed during the cue period. Figure 5B illustrates event-related cue-period activity recorded from the same 20 sensors as in Figure 5A. Ichihara-Takeda et al.
457
Figure 3. (A) TSE curves of alpha band oscillations in 204 sensors for a participant under the DMST (solid line) and DMST+d (dotted line) conditions. The visual cue was presented at t = 0 sec; the onset of the delay period, at t = 0.5 sec; the end signal, at t = 5.5 sec; and the probe, at t = 6 sec. The distractors in the DMST+d condition were presented every second from the onset of the delay period to the onset of end signal. Note the modulation of alpha activity in the occipital and parietal sensors in the DMST condition. (B) The selected occipital sensor was the most reactive in the DMST condition. The solid and dotted lines indicate the TSE curves obtained in the DMST and DMST+d conditions, respectively.
458
Journal of Cognitive Neuroscience
Volume 27, Number 3
Figure 4. Effects of different conditions on alpha activity in seven participants. The TSE curves were obtained from the most reactive sensor. Blue, light blue, red, and pink lines indicate the TSE curves obtained in the DMST, CONT, DMST+d, and CONT+d conditions, respectively. Ordinates: TSE% value based on the baseline activity. Abscissas: The visual cue was presented at t = 0 sec; the delay period began at t = 0.5 sec; the end signal, at t = 5.5 sec; and the probe, at t = 6 sec. Arrows indicate the onset of the first distractor.
The location of the most reactive sensor for the suppression of alpha activity during the cue period was located next to the sensor that exhibited the strongest event-related cue-period activity. The same spatial relationship was observed in all five participants who showed the suppression of alpha activity during the cue period in the DMST condition. Thus, the location of the sensor that showed the suppression of alpha activity during the cue period was
located next to the main location for the handling of lowlevel information for visual signals. Effect of Distractors on Alpha Activity in a DMST The TSE curves of alpha band oscillations shown in Figure 3 also show the effect of distractors. Although an increase of alpha activity during the delay period of the Ichihara-Takeda et al.
459
DMST condition was observed in parieto-occipital sensors, an increase of alpha activity was not observed in the DMST +d condition (Figure 3A). Figure 3B shows no increase of alpha activity during the delay period of the DMST+d condition, even in the most reactive sensor in the DMST condition. No increase of alpha activity was observed during the delay period of the DMST+d condition in all seven participants who showed an increase of alpha activity in the DMST condition (Figure 4). In Participants 1 and 2, no increase of alpha activity was observed during either the cue period or the delay period in the DMST+d condition compared with the resting level. However, in the remaining five participants (Participants 3–7), alpha activity
decreased to below the resting level by the end of the cue period and remained at this level throughout the delay period in the DMST+d condition. With regard to the timing of the difference in TSE curves between the DMST and DMST+d conditions, in four participants (Participants 1, 2, 4, and 7), this difference arose before the onset of the first distractor and was maintained until the end of the delay period. At the end-signal period, the difference of alpha activity rapidly diminished in all seven participants, and the strength of alpha activity was lowest in the first half of the probe period. The strength of alpha activity then rapidly increased toward the end of the probe period. The temporal pattern of alpha activity from the end-signal
Figure 5. (A) TSE waveforms of alpha band activity detected by 20 sensors around the most reactive sensor for the DMST condition (solid circle). The most reactive sensor was located in the occipital area. The solid and dotted lines indicate the TSE curves obtained for the DMST and CONT conditions, respectively. The onset of the visual cue was at t = 0 sec; the onset of the delay period, at t = 0.5 sec; the end signal, at t = 5.5 sec; and the probe, at t = 6 sec. (B) Corresponding event-related cue-period activity for the same electrodes as those shown in A for the same participant. The dotted circle indicates the sensor with the strongest event-related cue-period activity in the occipital area. Note that the most reactive sensor as shown by the TSE curve (solid circle) was located next to that with the strongest event-related cue-period activity (dotted circle) in the occipital area.
460
Journal of Cognitive Neuroscience
Volume 27, Number 3
Figure 6. Mean values of the TSE% levels of alpha activity across the seven participants who showed alpha activity modulation (each column, n = 7). The TSE level was obtained from the most reactive sensor.
period to the probe period in the DMST+d condition was similar to that observed in the DMST condition. The temporal pattern of alpha activity in the CONT+d condition was similar to that in the DMST+d condition in all seven participants. Thus, no increase of alpha activity with distractors was observed during the delay period in either the DMST+d or CONT+d condition. To quantify the effects of different conditions on the TSE level during the delay period, mean TSE% was calculated for the seven participants who exhibited temporal modulation of the strength of alpha activity (Figure 6). The mean TSE% was 126 in the DMST condition, 95 in the DMST+d condition, 115 in the CONT condition, and 97 in the CONT+d condition. Two-way repeated-measures ANOVA revealed a main effect for the presence or absence of distractor stimuli (without vs. with distractor conditions; F(1, 6) = 70.7, p < .001). We did not observe a significant effect for the task difference (DMST vs. CONT) or for interactions between tasks and distractor conditions.
DISCUSSION Does Alpha Activity in the Parieto-occipital Areas during the Visuospatial Working Memory Task Directly Reflect Working Memory Processes? We investigated the functions of alpha activity observed in the parieto-occipital areas during the visuospatial working memory task. We used two behavioral tasks (DMST and CONT) that did and did not require working memory, respectively, under two conditions without and with distractors. We analyzed cortical alpha activity by MEG. The largest temporal change of alpha activity was observed at sensors located in the parieto-occipital area during the DMST and CONT conditions. This temporal modulation was linked to the progress of the task, such that the strength of alpha activity increased after the onset of the delay period, was maintained during the
delay period, and rapidly weakened after the onset of the end-signal period in this study. Previous electroencephalography and MEG studies have also shown an increase of alpha activity in the parieto-occipital area during a working memory task (Sauseng et al., 2009; Ciesielski et al., 2007; Jokisch & Jensen, 2007; Tuladhar et al., 2007; Jensen et al., 2002). Jensen et al. (2002) showed that the modulation of alpha activity was linked to the working memory task such that the activity rapidly diminished as soon as the delay period ended. The tight temporal regulation of alpha activity in the parieto-occipital area strongly suggests a linkage between the alphagenerating system and the circuits responsible for the maintenance of information in working memory. However, the function of alpha activity in working memory tasks is open to several interpretations, including two leading hypotheses. One interpretation is that alpha activity directly reflects neuronal processing that is required for the maintenance of information in working memory (Palva & Palva, 2007; Palva et al., 2005). The other possibility is that the increase of alpha activity reflects functional inhibition or disengagement (Jensen & Mazaheri, 2010; Klimesch et al., 2007). In our study, a control task that did not require working memory (the CONT condition) was introduced. As a result, an increase of alpha activity was observed in the CONT condition in five of the seven participants who exhibited an increase of alpha activity during the delay period in the DMST condition. This is consistent with the findings of Jokisch and Jensen (2007), who showed that strong alpha activity was observed in a control task that did not require working memory. To further examine the functions of alpha activity in the parieto-occipital area, we introduced visuospatial distractors during the delay period of the visuospatial working memory task. TSE analysis revealed that there was no increase of alpha activity during the delay period in the DMST+d condition in any of the seven participants who exhibited an increase of alpha activity in the DMST condition, although TSE curves were constructed only from correct trials. These results support the idea that the increase of alpha activity detected by parieto-occipital sensors does not directly reflect an active neural process that is related to the maintenance of information in working memory. What Function Does Alpha Activity in the Parieto-occipital Area Reflect? Recent studies have suggested that the increase of alpha activity observed in the parieto-occipital area reflects the disengagement or inhibition of the visual dorsal stream ( Jensen & Mazaheri, 2010; Jokisch & Jensen, 2007; Klimesch et al., 2007; Tuladhar et al., 2007). A systematic increase of alpha activity as the memory load increased was observed over posterior brain areas during the delay period. This phenomenon has been considered to reflect the disengagement or inhibition. Such disengagement was thought to suppress visual input so that necessary Ichihara-Takeda et al.
461
resources could be devoted to brain structures responsible for maintaining working memory (Tuladhar et al., 2007). In our study, although the difference was not statistically significant, the percentage of correct performance in the DMST+d condition was lower than that in the DMST condition. Therefore, if we can assume that the memory load was greater in the DMST+d condition than in the DMST condition, it is possible that alpha activity in the parietooccipital area increased more in the DMST+d condition than in the DMST condition. However, our present results showed that alpha activity was suppressed during the delay period in both the DMST+d and CONT+d conditions. Previous studies have shown that visual attention modulated the strength of oscillatory alpha activity (Rihs, Michel, & Thut, 2007; Sauseng et al., 2005; Foxe, Simpson, & Ahlfors, 1998). An increase of alpha activity in the parieto-occipital area emerged when participants rested with their eyes closed (Salenius, Kajola, Thompson, Kosslyn, & Hari, 1995). Because the suppression of alpha activity was greater after the presentation of meaningful objects than after the presentation of meaningless nonobjects, it was thought that the difference of alpha activity was related to the strength of attention to the objects required by the task (Vanni, Revonsuo, & Hari, 1997). Ciesielski et al. (2007) showed that patients with obsessive compulsive disorder could maintain accurate performance despite an abnormally low alpha modulation in the parietooccipital area during the working memory task. Several studies have indicated that the strength of alpha activity is affected by the anticipatory attention effect (Haegens, Luther, & Jensen, 2012; Sauseng et al., 2005; Worden et al., 2000; Foxe et al., 1998). Our present results showed that alpha activity was suppressed not only by the presentation of distractors but also by the presentation of the visual cue, the end signal, the probe stimuli, and Japanese characters. We also found that, in some participants, alpha activity was suppressed before the distractor was presented. Thus, we can assume that alpha activity may be suppressed by the presentation or anticipation of the visual stimuli to which the participant needed to attend and an increase of alpha activity may reflect the disengagement of attention from the visual input. The present study showed that the increases of alpha activity occurred regardless of whether working memory was required. This suggests that alpha activity reflects disengagement rather than memory maintenance. The presentation of distractors, which had spatial information, reduced alpha activity. This can be explained by the notion that the presentation or anticipation of visual stimuli increased attention or reduced the disengagement. Thus, the present result supports the interpretation that the increase of alpha activity reflects the disengagement. In addition, although the suppression of alpha activity by the presentation of Japanese characters was transient, the suppression by the presentation of distractors was maintained. In this study, Japanese characters were presented at the center of the screen, whereas distractors were 462
Journal of Cognitive Neuroscience
presented around the characters and thus had spatial information. Alpha activity may be suppressed by spatial information. Therefore, we could conclude that the increase of alpha activity in the parieto-occipital area reflects the disengagement of attention from the visuospatial input. Acknowledgments The authors thank Dr. K. Takeda for offering valuable comments regarding the working memory task, Mr. A. Yoshida for providing technical support for the working memory task, and Ms. E. Takeda for providing support for MEG recording. This study was supported by a Grant-in-Aid for Young Scientists (B; 21700541, 24700542) to S. T. and a Grant-in-Aid for Scientific Research (C; 22500373) to T. N. from the Japan Society for the Promotion of Science. Reprint requests should be sent to Dr. Satoe Ichihara-Takeda, Department of Occupational Therapy, School of Health Science, Sapporo Medical University, South 1 West 17, Chuo-ku, Sapporo 060-8556, Japan, or via e-mail: [email protected].
REFERENCES Baddeley, A. (1986). Working memory. Oxford: Oxford University Press. Carlson, S., Martinkauppi, S., Rämä, P., Salli, E., Korvenoja, A., & Aronen, H. J. (1998). Distribution of cortical activation during visuospatial n-back tasks as revealed by functional magnetic resonance imaging. Cerebral Cortex, 8, 743–752. Chafee, M. V., & Goldman-Rakic, P. S. (1998). Matching patterns of activity in primate prefrontal area 8a and parietal area 7ip neurons during a spatial working memory task. Journal of Neurophysiology, 79, 2919–2940. Ciesielski, K. T., Hämäläinen, M. S., Geller, D. A., Wilhelm, S., Goldsmith, T. E., & Ahlfors, S. P. (2007). Dissociation between MEG alpha modulation and performance accuracy on visual working memory task in obsessive compulsive disorder. Human Brain Mapping, 28, 1401–1414. Curtis, C. E., & DʼEsposito, M. (2003). Persistent activity in the prefrontal cortex during working memory. Trends in Cognitive Science, 7, 415–423. Foxe, J. J., Simpson, G. V., & Ahlfors, S. P. (1998). Parieto-occipital approximately 10 Hz activity reflects anticipatory state of visual attention mechanisms. NeuroReport, 9, 3929–3933. Funahashi, S., Bruce, C. J., & Goldman-Rakic, P. S. (1989). Mnemonic coding of visual space in the monkeyʼs dorsolateral prefrontal cortex. Journal of Neurophysiology, 61, 331–349. Goldman-Rakic, P. S. (1987). Circuitry of primate prefrontal cortex and regulation of behavior by representational memory. In F. Plum (Ed.), Handbook of physiology, sect. 1, vol. V: The nervous system. Higher functions of the brain (pp. 373–417). Bethesda, MD: American Physiological Society. Haegens, S., Luther, L., & Jensen, O. (2012). Somatosensory anticipatory alpha activity increases to suppress distracting input. Journal of Cognitive Neuroscience, 24, 677–685. Ichihara-Takeda, S., & Funahashi, S. (2007). Activity of primate orbitofrontal and dorsolateral prefrontal neurons: Task-related activity during an oculomotor delayed-response task. Experimental Brain Research, 181, 409–425. Ichikawa, A., Yamamoto, H., Ono, I., Matsubayashi, J., Nagamine, T., Fukuyama, H., et al. (2007). Stimulus-related 20-Hz activity of human cortex modulated by the way of presenting hand actions. Neuroscience Research, 58, 285–290.
Volume 27, Number 3
Jensen, O., Gelfand, J., Kounios, J., & Lisman, J. E. (2002). Oscillations in the alpha band (9–12 Hz) increase with memory load during retention in a short-term memory task. Cerebral Cortex, 12, 877–882. Jensen, O., & Mazaheri, A. (2010). Shaping functional architecture by oscillatory alpha activity: Gating by inhibition. Frontiers in Human Neuroscience, 4, 186. Jokisch, D., & Jensen, O. (2007). Modulation of gamma and alpha activity during a working memory task engaging the dorsal or ventral stream. Journal of Neuroscience, 27, 3244–3251. Klimesch, W., Sauseng, P., & Hanslmayr, S. (2007). EEG alpha oscillations: The inhibition-timing hypothesis. Brain Research Reviews, 53, 63–88. Nagamine, T., Kajola, M., Salmelin, R., Shibasaki, H., & Hari, R. (1996). Movement-related slow cortical magnetic fields and changes of spontaneous MEG- and EEG-brain rhythms. Electroencephalography and Clinical Neurophysiology, 99, 274–286. Palva, J. M., Palva, S., & Kaila, K. (2005). Phase synchrony among neuronal oscillations in the human cortex. Journal of Neuroscience, 25, 3962–3972. Palva, S., & Palva, J. M. (2007). New vistas for alpha-frequency band oscillations. Trends in Neuroscience, 30, 150–158. Park, S., & Holzman, P. S. (1992). Schizophrenics show spatial working memory deficits. Archives of General Psychiatry, 49, 975–982. Park, S., Holzman, P. S., & Goldman-Rakic, P. S. (1995). Spatial working memory deficits in the relatives of schizophrenic patients. Archives of General Psychiatry, 52, 821–828. Pfurtscheller, G., Stancák, A., Jr., & Neuper, C. (1996). Event-related synchronization (ERS) in the alpha band—An electrophysiological correlate of cortical idling: A review. International Journal of Psychophysiology, 24, 39–46. Postle, B. R., & DʼEsposito, M. (1999). Dissociation of human caudate nucleus activity in spatial and nonspatial working memory: An event-related fMRI study. Brain Research: Cognitive Brain Research, 8, 107–115. Rihs, T. A., Michel, C. M., & Thut, G. (2007). Mechanisms of selective inhibition in visual spatial attention are indexed by alpha-band EEG synchronization. European Journal of Neuroscience, 25, 603–610. Roesch, M. R., & Olson, C. R. (2005). Neuronal activity dependent on anticipated and elapsed delay in macaque prefrontal cortex, frontal and supplementary eye fields, and premotor cortex. Journal of Neurophysiology, 94, 1469–1497. Sakai, K., Rowe, J. B., & Passingham, R. E. (2002). Active maintenance in prefrontal area 46 creates distractor-resistant memory. Nature Neuroscience, 5, 479–484.
Salenius, S., Kajola, M., Thompson, W. L., Kosslyn, S., & Hari, R. (1995). Reactivity of magnetic parieto-occipital alpha rhythm during visual imagery. Electroencephalography and Clinical Neurophysiology, 95, 453–462. Salmelin, R., Hämäläinen, M., Kajola, M., & Hari, R. (1995). Functional segregation of movement-related rhythmic activity in the human brain. Neuroimage, 2, 237–243. Salmelin, R., & Hari, R. (1994). Spatiotemporal characteristics of sensorimotor neuromagnetic rhythms related to thumb movement. Neuroscience, 60, 537–550. Sauseng, P., Klimesch, W., Heise, K. F., Gruber, W. R., Holz, E., Karim, A. A., et al. (2009). Brain oscillatory substrates of visual short-term memory capacity. Current Biology, 19, 1846–1852. Sauseng, P., Klimesch, W., Stadler, W., Schabus, M., Doppelmayr, M., Hanslmayr, S., et al. (2005). A shift of visual spatial attention is selectively associated with human EEG alpha activity. European Journal of Neuroscience, 22, 2917–2926. Schnitzler, A., Salenius, S., Salmelin, R., Jousmäki, V., & Hari, R. (1997). Involvement of primary motor cortex in motor imagery: A neuromagnetic study. Neuroimage, 6, 201–208. Tamura, Y., Hoshiyama, M., Nakata, H., Hiroe, N., Inui, K., Kaneoke, Y., et al. (2005). Functional relationship between human rolandic oscillations and motor cortical excitability: An MEG study. European Journal of Neuroscience, 21, 2555–2562. Thut, G., Nietzel, A., Brandt, S. A., & Pascual-Leone, A. (2006). Alpha-band electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts visual target detection. Journal of Neuroscience, 26, 9494–9502. Tuladhar, A. M., ter Huurne, N., Schoffelen, J. M., Maris, E., Oostenveld, R., & Jensen, O. (2007). Parieto-occipital sources account for the increase in alpha activity with working memory load. Human Brain Mapping, 28, 785–792. Vanni, S., Revonsuo, A., & Hari, R. (1997). Modulation of the parieto-occipital alpha rhythm during object detection. Journal of Neuroscience, 17, 7141–7147. Watanabe, Y., & Funahashi, S. (2004). Neuronal activity throughout the primate mediodorsal nucleus of the thalamus during oculomotor delayed-responses: I. Cue-, delay-, and response-period activity. Journal of Neurophysiology, 92, 1738–1755. Worden, M. S., Foxe, J. J., Wang, N., & Simpson, G. V. (2000). Anticipatory biasing of visuospatial attention indexed by retinotopically specific alpha-band electroencephalography increases over occipital cortex. Journal of Neuroscience, 20, RC63.
Ichihara-Takeda et al.
463
Copyright of Journal of Cognitive Neuroscience is the property of MIT Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Abstract ■ Oscillatory brain activity is known to play an essential role in
information processing in working memory. Recent studies have indicated that alpha activity (8–13 Hz) in the parieto-occipital area is strongly modulated in working memory tasks. However, the function of alpha activity in working memory is open to several interpretations, such that alpha activity may be a direct neural correlate of information processing in working memory or may reflect disengagement from information processing in other brain areas. To examine the functional contribution of alpha activity to visuospatial working memory, we introduced visuospatial distractors during a delay period and examined neural activity from the whole brain using magnetoencephalography.
INTRODUCTION Working memory is a mechanism for the short-term active storage of information as well as for the processing of stored information (Baddeley, 1986). Goldman-Rakic (1987) proposed that working memory is an important concept for understanding dorsolateral prefrontal cortex (DLPFC) functions. Single-cell recordings and functional MRI studies have demonstrated that excitatory sustained activity is observed in the DLPFC during the delay period in the visuospatial working memory task (Ichihara-Takeda & Funahashi, 2007; Roesch & Olson, 2005; Curtis & DʼEsposito, 2003; Sakai, Rowe, & Passingham, 2002; Funahashi, Bruce, & Goldman-Rakic, 1989). Therefore, this sustained activity has been considered to be a neural correlate of the temporary maintenance of information. However, this sustained activity is observed not only in the DLPFC but also in several other brain areas, including the posterior parietal cortex (Sakai et al., 2002; Chafee & Goldman-Rakic, 1998), occipital visual association areas (Carlson et al., 1998), the caudate nucleus (Postle & DʼEsposito, 1999), and the
1
Sapporo Medical University, Japan, 2Hokkaido University School of Medicine, Sapporo, Japan, 3Kyoto University © 2015 Massachusetts Institute of Technology
The strength of event-related alpha activity was estimated using the temporal spectral evolution (TSE) method. The results were as follows: (1) an increase of alpha activity during the delay period as indicated by elevated TSE curves was observed in parietooccipital sensors in both the working memory task and a control task that did not require working memory; and (2) an increase of alpha activity during the delay period was not observed when distractors were presented, although TSE curves were constructed only from correct trials. These results indicate that the increase of alpha activity is not directly related to information processing in working memory but rather reflects the disengagement of attention from the visuospatial input. ■
thalamic mediodorsal nucleus ( Watanabe & Funahashi, 2004). These results indicate that several brain areas participate in working memory. Therefore, to better understand the neural mechanisms of working memory, it is important to measure neuronal activity from the whole brain with high temporal resolution while participants perform working memory tasks. Electroencephalography and magnetoencephalography (MEG) can be used to measure whole-brain neuronal activity with high temporal resolution. Studies that have used these techniques suggest that oscillatory alpha activity in posterior cortical areas reflects an idling condition of the cortex (Pfurtscheller, Stancák, & Neuper, 1996) or a preparatory process for attention (Thut, Nietzel, Brandt, & Pascual-Leone, 2006; Worden, Foxe, Wang, & Simpson, 2000). Oscillatory alpha activity has also been analyzed to determine the neuronal processes that are necessary for working memory. Alpha activity is strongly modulated during working memory tasks, and this activity is often observed in the parieto-occipital area in addition to the prefrontal area (Ciesielski et al., 2007; Jokisch & Jensen, 2007; Tuladhar et al., 2007; Jensen, Gelfand, Kounios, & Lisman, 2002). Especially, the increase of alpha activity in the parieto-occipital area is thought to reflect either an Journal of Cognitive Neuroscience 27:3, pp. 453–463 doi:10.1162/jocn_a_00718
active neural process related to memory maintenance (Palva & Palva, 2007; Palva, Palva, & Kaila, 2005) or the active inhibition of posterior cortical regions that are not required for the task (Jensen & Mazaheri, 2010; Klimesch, Sauseng, & Hanslmayr, 2007). Recent studies have suggested that parieto-occipital alpha activity during the delay period of working memory tasks reflects the disengagement or inhibition of the visual dorsal stream (Jokisch & Jensen, 2007; Tuladhar et al., 2007). Individual differences of alpha activity appear to correlate with individual differences of the capacity for suppressing irrelevant information (Sauseng et al., 2009). However, the contribution of parieto-occipital alpha activity to visuospatial working memory processes continues to be debated. Therefore, in this study, we examined the signals or factors that affect the strength of alpha activity observed during the delay period of a visuospatial working memory task using MEG. We analyzed the temporal pattern of the strength of alpha activity in the whole brain in both working memory tasks and control tasks that did not require working memory. In addition, we introduced visuospatial distractors during the delay period of the task to estimate the effect of alpha activity on the maintenance of information in working memory.
METHODS Participants The experiments were carried out in 10 healthy participants (seven men, nine right-handed and one left-handed, age range = 21–49 years, mean = 30.9 years). Informed consent was obtained from each participant before the study. The study protocol was approved by the ethics committees of Sapporo Medical University and Hokkaido University. Task In this study, we used two tasks (a delayed matching-tosample task [DMST] and a control task [CONT]) under two distractor conditions (without and with distractors). All participants performed the tasks under four conditions (each task without and with distractors: DMST and DMST +d, CONT and CONT+d). Trials under these four conditions consisted of five periods: fixation period (3 sec), cue period (0.5 sec), delay period (5 sec), end-signal period (0.5 sec), and probe period (0.8 sec). Visual stimuli were presented on a back-projection screen placed 1.4 m from the participant by an LCD projector (LT260SK; NEC, Tokyo, Japan). Each trial started with the presentation of Japanese characters (hiragana; white, 1.5° in visual angle). The characters were presented at the center of the screen to enforce the participantsʼ maintenance of eye fixation at the center of the screen and to prevent the use of language to maintain spatial information of the visual cue during the delay 454
Journal of Cognitive Neuroscience
period (Park, Holzman, & Goldman-Rakic, 1995; Park & Holzman, 1992). Participants were required to view the Japanese characters and read them silently throughout the fixation, cue, and delay periods (Figure 1). The Japanese characters were randomly chosen from a pool of 50 characters and presented successively, one by one, for 1 sec each. After the 3-sec fixation period, a visual cue (white circle, 2.5° in visual angle) was presented for 0.5 sec at a peripheral location simultaneously with the presentation of a Japanese character at the center (cue period). The location of the visual cue was randomly selected from eight predetermined positions. The 5-sec delay period then began. At the end of the delay period, the central Japanese character was replaced with a star figure, which signaled the end of the delay period, and remained on the screen for 0.5 sec (end-signal period). After the end-signal period, a probe cue was presented for 0.8 sec (probe period). In the DMST condition, participants were required to make a same–different judgment based on whether the visual cue and the probe cue appeared at the same position by lifting either the right index finger (same position) or the left index finger (different position; Figure 1A). The visual cue and the probe cue appeared at the same location in 12.5% of trials. In the CONT condition, participants were required to lift the left index finger when the probe cue was presented, regardless of its location. Therefore, they did not need to remember the position of the visual cue. After the participants responded, an intertrial interval of 3 sec was introduced. During the intertrial interval, participants were allowed to blink. In the DMST with distractor condition (DMST+d), distractors were presented during the delay period. Each distractor (white circle, 2.5° in visual angle) was presented for 0.5 sec at a location that was randomly selected from among the eight predetermined positions. The first distractor was presented 1 sec after the start of the delay period, and three other distractors were presented sequentially every 1 sec during the delay period. Their color and shape were the same as those of the visual cue (Figure 1B). In the CONT with distractor condition (CONT+d), participants did not need to remember the position of the visual cue, and distractors were presented according to the same procedure as in the DMST+d condition. The participants were exposed to four conditions (DMST, CONT, DMST+d, and CONT+d) on 2 separate days, and the order was randomized among participants. On the first day, participants performed two different conditions twice, for example, DMST (first session), DMST+d (first session), DMST (second session), and DMST+d (second session). Each session included 60 trials. On the second day, the participants performed the remaining two conditions twice, for example, CONT+d (first session), CONT (first session), CONT+d (second session), and CONT (second session). Volume 27, Number 3
Figure 1. The experimental paradigm. Two variants of a modified DMST were used: DMST with no distractor (DMST, A) and with distractors (DMST+d, B). (A) Each trial began with the presentation of Japanese characters. From the start of the fixation period until the end signal, the Japanese characters (hiragana) were presented at the center of the screen to help the participant maintain eye fixation at the center of the screen and to prevent the use of language to maintain spatial information during the delay period. After the 3-sec fixation period, a visual cue was presented for 0.5 sec (cue period). The location of the visual cue was randomly selected from eight predetermined positions. The 5-sec delay period (delay period) then began. At the end of the delay period, the central Japanese character was replaced with a star figure, which remained on the screen for 0.5 sec (end-signal period). After the end-signal period, a visual probe stimulus was presented for 0.8 sec (probe period). Participants were requested to make a same–different judgment by lifting their right index finger for “same” if the probe appeared at the same position as the cue. Participants were also requested to make a judgment by lifting their left index finger for “different” if the probe appeared in a different position than the cue. (B) In the DMST+d condition, distractors were presented during the delay period. Each distractor was presented for 0.5 sec at a location that was randomly selected from eight predetermined positions. The first distractor was presented 1 sec after the start of the delay period, and three other distractors were presented sequentially every 1 sec during the delay period. The color and shape of the distractors were the same as those of the visual cue. (C) Timing and display area for the presentation of Japanese characters (hiragana), cue, end signal, probe, and distractors during the task.
Ichihara-Takeda et al.
455
MEG Recordings We used MEG with whole-head 204-channel planar gradiometers (Vectorview; Elekta, Stockholm, Sweden) in a magnetically shielded room. The sampling frequency was 600 Hz, and the bandpass filter was set at 0.03–200 Hz. Eye movements and eye blinks were monitored by a vertical EOG. The participantsʼ motor performance (extension of the index finger) was continuously monitored by an online electromyogram recorded from the extensor indicis muscle. Anatomical landmark positions (nasion and preauricular points) were determined by a Polhemus Fastrak 3-D digitizer (Colchester, Vermont). Because the sessions were separated by 2- to 3-min breaks, each participantʼs head position relative to the MEG sensor array was measured at the beginning of each session. Before recording on the second day, each participantʼs head position relative to the MEG sensor array was measured, and the position of their head was adjusted so that the displacement of head position from the first recording was within 10 mm.
et al., 2005; Schnitzler, Salenius, Salmelin, Jousmäki, & Hari, 1997). The following time windows were analyzed: (1) −2 to 0 sec with respect to the onset of the cue period (baseline period), (2) 0–0.5 sec after the onset of the cue period (cue period), and (3) 0.5–5.5 sec after the onset of the cue period (delay period). For the purposes of this study, the delay period was of particular interest. The mean values of TSE levels over the baseline and delay periods were calculated for the most reactive sensor of the gradiometer. TSE% was defined as the percentage increase in band power during the delay period as compared with the baseline period: TSE% = (band power, delay period)/(band power, baseline period) × 100. This value was compared across four conditions using twoway repeated-measures ANOVA with two tasks (DMST and CONT) and two kinds of stimuli (without and with distractors) as factors.
RESULTS Data Analysis Epochs from 2 sec before the onset of the cue period to 6.8 sec after the onset of the cue period were extracted from continuous MEG raw data. The activity during each epoch was visually inspected, and trials that included large eye movements or blinking were excluded from the analysis. Trials in which participants made incorrect judgments were also excluded from the analysis. The level of event-related alpha activity as a function of time was estimated using the temporal spectral evolution (TSE) method (Nagamine, Kajola, Salmelin, Shibasaki, & Hari, 1996; Salmelin, Hämäläinen, Kajola, & Hari, 1995; Salmelin & Hari, 1994). TSE analysis was adopted to characterize the reactivity of the spontaneous cortical activity. In this analysis, spontaneous activity was extracted by using the frequency filter of interest. Thereafter, its temporal profile with respect to events of stimulus or movement onset was obtained by averaging the absolute value of filtered data across events. In our data analysis, the continuous MEG signals were filtered through pass bands of 8–13 Hz and then rectified. The absolute values within the selected epochs were averaged over the trials with respect to the onset of the cue period. Because we were mainly interested in the slow component with a temporal resolution of 100 msec, we smoothed the data with a 10-Hz low-pass filter. In this study, approximately 100–120 trials in each condition were averaged with respect to the onset of the cue period. However, only 60 trials were averaged for Participant 7 because of the frequent occurrence of artifacts during epochs used for analysis. Because MEG signals from planar gradiometers are strongest when the sensors are located above cortical current sources, the data from the most reactive sensor that showed the strongest TSE response were used to evaluate the effects of various experimental conditions (Ichikawa et al., 2007; Tamura 456
Journal of Cognitive Neuroscience
Behavioral Data To examine whether behavioral performance was affected by distractors during the delay period, the percentage of correct performance was calculated for each participant. Figure 2 shows the percentage of correct performance in the DMST and DMST+d conditions for all 10 participants. In five participants, the percentage of correct performance in the DMST+d condition was lower than that in the DMST condition. Three participants showed the same percentage in these two conditions, and two showed a slightly higher percentage in the DMST+d condition than in the DMST condition. The mean correct percentage among the 10 participants was 97.4% (SD = 2.3) for the
Figure 2. Behavioral data during MEG sessions. The percentage of correct performance in the DMST and DMST+d conditions is plotted for all 10 participants.
Volume 27, Number 3
DMST condition and 94.2% (SD = 7.3) for the DMST+d condition. Although the overall performance in the DMST condition was slightly better than that in the DMST+d condition, this difference was not significant (Wilcoxon ranksum test, p = .18). Alpha Activity in the DMST and the Control Task Figure 3A shows a representative participantʼs TSE curves for alpha-band oscillations constructed from the recordings of 204 sensors. The TSE curve that showed the largest change in alpha activity was observed for the parieto-occipital sensors during the DMST condition. Figure 3B shows TSE curves obtained from the most reactive sensor that showed the strongest alpha activity. Temporal modulation was clearly observed during the delay period in the DMST condition. The activity increased approximately 1 sec after the onset of the cue period (0.5 sec after the onset of the delay period). The activity peaked at 3.3 sec after the onset of the cue period and then gradually became weaker. At the end-signal period, the activity rapidly diminished and reached its lowest level in the first half of the probe period, which was lower than that during the baseline period. The activity then rapidly increased starting 0.4 sec after presentation of the probe and continued to increase beyond the end of the probe period. We observed similar temporal patterns of alpha activity during the delay period in the DMST condition in 7 of the 10 participants. However, in the remaining three participants, the modulation of alpha activity was not observed during any task epoch at any cortical location in any of the conditions. The percentage of correct performance during the DMST condition in the seven participants who showed modulation was 97.4% (SD = 2.3) and that in the three participants who did not show modulation was 97.3% (SD = 2.7). Thus, there was no difference in the correct performance rate between participants who did and did not show an increase of alpha activity during the delay period of the DMST condition. Figure 4 shows TSE curves obtained from the most reactive sensor that showed the strongest alpha activity in these seven participants. In all of the participants, these sensors were located in the parieto-occipital area. Alpha activity started to increase after the initiation of the delay period in the DMST condition in Participants 1 and 2. In the remaining five participants (Participants 3–7), alpha activity was first suppressed during the cue period and then increased during the delay period in the DMST condition. An increase of alpha activity similar to the increase of activity during the delay period in the DMST condition was observed in the CONT condition (Participants 2, 4, 5, 6, and 7). The alpha activity in the CONT condition as a percentage of that in the DMST condition in Participants 2, 4, 5, 6, and 7 was 95.1%, 90.7%, 98.3%, 96.6%, and 103%, respectively. However, in Participants 1 and 3,
the alpha activity in the CONT condition as a percentage of that in the DMST condition was 78.2% and 79.2%, respectively. Thus, alpha activity during the delay period increased in the CONT condition as much as in the DMST condition in these five participants. The suppression of alpha activity during the cue period was also observed in the CONT condition in all participants who showed similar suppression in the DMST condition. Furthermore, in both the DMST and CONT conditions, all seven participants demonstrated the rapid suppression of alpha activity after the onset of the endsignal period. The alpha activity reached the lowest value approximately 0.3 sec after the onset of the probe period. This activity then started increasing rapidly approximately 0.4 sec after the onset of the probe period. In addition, the transient suppression of increased alpha activity was observed from 2, 3, 4, and 5 sec after the onset of the cue stimulus in six participants in the DMST and CONT conditions. In this task, Japanese characters were presented successively, one by one, for 1 sec each. The timing of the transient suppression of alpha activity corresponded to that when a Japanese character was presented. Thus, the suppression of activity coincided with the presentation of visual stimuli (cue, end signal, probe, and Japanese characters). We should also stress that this suppression was observed not only during the visuospatial working memory task but also during the control task that did not require working memory. On the other hand, the temporal pattern of the increase of alpha activity after the suppression caused by cue presentation was different from that caused by presentation of the probe stimulus. After the suppression caused by cue presentation, it took about 1.5 sec for alpha activity to return to the baseline level. On the other hand, a rapid increase of activity was observed in the middle of the probe period. A Japanese character was presented 1 sec after the onset of the cue stimulus (0.5 sec after the onset of the delay period), whereas no other visual stimulus was presented after the onset of the probe stimulus. Thus, the timing of the slow increase of alpha activity after the suppression caused by cue presentation corresponded to that when a Japanese character was presented. In the DMST condition, five of seven participants exhibited the suppression of alpha activity during the cue period. To identify the brain regions that were responsible for this suppression of alpha activity during the cue period, the cortical distribution was compared between the suppression of alpha activity as shown by TSE curves and event-related cue period activity. Twenty sensors were selected for this comparison from among sensors adjacent to the sensor that exhibited the strongest alpha activity according to TSE curves in the DMST condition (Figure 5). Figure 5A illustrates examples of TSE curves obtained from sensors within the parieto-occipital area where the suppression of alpha activity was observed during the cue period. Figure 5B illustrates event-related cue-period activity recorded from the same 20 sensors as in Figure 5A. Ichihara-Takeda et al.
457
Figure 3. (A) TSE curves of alpha band oscillations in 204 sensors for a participant under the DMST (solid line) and DMST+d (dotted line) conditions. The visual cue was presented at t = 0 sec; the onset of the delay period, at t = 0.5 sec; the end signal, at t = 5.5 sec; and the probe, at t = 6 sec. The distractors in the DMST+d condition were presented every second from the onset of the delay period to the onset of end signal. Note the modulation of alpha activity in the occipital and parietal sensors in the DMST condition. (B) The selected occipital sensor was the most reactive in the DMST condition. The solid and dotted lines indicate the TSE curves obtained in the DMST and DMST+d conditions, respectively.
458
Journal of Cognitive Neuroscience
Volume 27, Number 3
Figure 4. Effects of different conditions on alpha activity in seven participants. The TSE curves were obtained from the most reactive sensor. Blue, light blue, red, and pink lines indicate the TSE curves obtained in the DMST, CONT, DMST+d, and CONT+d conditions, respectively. Ordinates: TSE% value based on the baseline activity. Abscissas: The visual cue was presented at t = 0 sec; the delay period began at t = 0.5 sec; the end signal, at t = 5.5 sec; and the probe, at t = 6 sec. Arrows indicate the onset of the first distractor.
The location of the most reactive sensor for the suppression of alpha activity during the cue period was located next to the sensor that exhibited the strongest event-related cue-period activity. The same spatial relationship was observed in all five participants who showed the suppression of alpha activity during the cue period in the DMST condition. Thus, the location of the sensor that showed the suppression of alpha activity during the cue period was
located next to the main location for the handling of lowlevel information for visual signals. Effect of Distractors on Alpha Activity in a DMST The TSE curves of alpha band oscillations shown in Figure 3 also show the effect of distractors. Although an increase of alpha activity during the delay period of the Ichihara-Takeda et al.
459
DMST condition was observed in parieto-occipital sensors, an increase of alpha activity was not observed in the DMST +d condition (Figure 3A). Figure 3B shows no increase of alpha activity during the delay period of the DMST+d condition, even in the most reactive sensor in the DMST condition. No increase of alpha activity was observed during the delay period of the DMST+d condition in all seven participants who showed an increase of alpha activity in the DMST condition (Figure 4). In Participants 1 and 2, no increase of alpha activity was observed during either the cue period or the delay period in the DMST+d condition compared with the resting level. However, in the remaining five participants (Participants 3–7), alpha activity
decreased to below the resting level by the end of the cue period and remained at this level throughout the delay period in the DMST+d condition. With regard to the timing of the difference in TSE curves between the DMST and DMST+d conditions, in four participants (Participants 1, 2, 4, and 7), this difference arose before the onset of the first distractor and was maintained until the end of the delay period. At the end-signal period, the difference of alpha activity rapidly diminished in all seven participants, and the strength of alpha activity was lowest in the first half of the probe period. The strength of alpha activity then rapidly increased toward the end of the probe period. The temporal pattern of alpha activity from the end-signal
Figure 5. (A) TSE waveforms of alpha band activity detected by 20 sensors around the most reactive sensor for the DMST condition (solid circle). The most reactive sensor was located in the occipital area. The solid and dotted lines indicate the TSE curves obtained for the DMST and CONT conditions, respectively. The onset of the visual cue was at t = 0 sec; the onset of the delay period, at t = 0.5 sec; the end signal, at t = 5.5 sec; and the probe, at t = 6 sec. (B) Corresponding event-related cue-period activity for the same electrodes as those shown in A for the same participant. The dotted circle indicates the sensor with the strongest event-related cue-period activity in the occipital area. Note that the most reactive sensor as shown by the TSE curve (solid circle) was located next to that with the strongest event-related cue-period activity (dotted circle) in the occipital area.
460
Journal of Cognitive Neuroscience
Volume 27, Number 3
Figure 6. Mean values of the TSE% levels of alpha activity across the seven participants who showed alpha activity modulation (each column, n = 7). The TSE level was obtained from the most reactive sensor.
period to the probe period in the DMST+d condition was similar to that observed in the DMST condition. The temporal pattern of alpha activity in the CONT+d condition was similar to that in the DMST+d condition in all seven participants. Thus, no increase of alpha activity with distractors was observed during the delay period in either the DMST+d or CONT+d condition. To quantify the effects of different conditions on the TSE level during the delay period, mean TSE% was calculated for the seven participants who exhibited temporal modulation of the strength of alpha activity (Figure 6). The mean TSE% was 126 in the DMST condition, 95 in the DMST+d condition, 115 in the CONT condition, and 97 in the CONT+d condition. Two-way repeated-measures ANOVA revealed a main effect for the presence or absence of distractor stimuli (without vs. with distractor conditions; F(1, 6) = 70.7, p < .001). We did not observe a significant effect for the task difference (DMST vs. CONT) or for interactions between tasks and distractor conditions.
DISCUSSION Does Alpha Activity in the Parieto-occipital Areas during the Visuospatial Working Memory Task Directly Reflect Working Memory Processes? We investigated the functions of alpha activity observed in the parieto-occipital areas during the visuospatial working memory task. We used two behavioral tasks (DMST and CONT) that did and did not require working memory, respectively, under two conditions without and with distractors. We analyzed cortical alpha activity by MEG. The largest temporal change of alpha activity was observed at sensors located in the parieto-occipital area during the DMST and CONT conditions. This temporal modulation was linked to the progress of the task, such that the strength of alpha activity increased after the onset of the delay period, was maintained during the
delay period, and rapidly weakened after the onset of the end-signal period in this study. Previous electroencephalography and MEG studies have also shown an increase of alpha activity in the parieto-occipital area during a working memory task (Sauseng et al., 2009; Ciesielski et al., 2007; Jokisch & Jensen, 2007; Tuladhar et al., 2007; Jensen et al., 2002). Jensen et al. (2002) showed that the modulation of alpha activity was linked to the working memory task such that the activity rapidly diminished as soon as the delay period ended. The tight temporal regulation of alpha activity in the parieto-occipital area strongly suggests a linkage between the alphagenerating system and the circuits responsible for the maintenance of information in working memory. However, the function of alpha activity in working memory tasks is open to several interpretations, including two leading hypotheses. One interpretation is that alpha activity directly reflects neuronal processing that is required for the maintenance of information in working memory (Palva & Palva, 2007; Palva et al., 2005). The other possibility is that the increase of alpha activity reflects functional inhibition or disengagement (Jensen & Mazaheri, 2010; Klimesch et al., 2007). In our study, a control task that did not require working memory (the CONT condition) was introduced. As a result, an increase of alpha activity was observed in the CONT condition in five of the seven participants who exhibited an increase of alpha activity during the delay period in the DMST condition. This is consistent with the findings of Jokisch and Jensen (2007), who showed that strong alpha activity was observed in a control task that did not require working memory. To further examine the functions of alpha activity in the parieto-occipital area, we introduced visuospatial distractors during the delay period of the visuospatial working memory task. TSE analysis revealed that there was no increase of alpha activity during the delay period in the DMST+d condition in any of the seven participants who exhibited an increase of alpha activity in the DMST condition, although TSE curves were constructed only from correct trials. These results support the idea that the increase of alpha activity detected by parieto-occipital sensors does not directly reflect an active neural process that is related to the maintenance of information in working memory. What Function Does Alpha Activity in the Parieto-occipital Area Reflect? Recent studies have suggested that the increase of alpha activity observed in the parieto-occipital area reflects the disengagement or inhibition of the visual dorsal stream ( Jensen & Mazaheri, 2010; Jokisch & Jensen, 2007; Klimesch et al., 2007; Tuladhar et al., 2007). A systematic increase of alpha activity as the memory load increased was observed over posterior brain areas during the delay period. This phenomenon has been considered to reflect the disengagement or inhibition. Such disengagement was thought to suppress visual input so that necessary Ichihara-Takeda et al.
461
resources could be devoted to brain structures responsible for maintaining working memory (Tuladhar et al., 2007). In our study, although the difference was not statistically significant, the percentage of correct performance in the DMST+d condition was lower than that in the DMST condition. Therefore, if we can assume that the memory load was greater in the DMST+d condition than in the DMST condition, it is possible that alpha activity in the parietooccipital area increased more in the DMST+d condition than in the DMST condition. However, our present results showed that alpha activity was suppressed during the delay period in both the DMST+d and CONT+d conditions. Previous studies have shown that visual attention modulated the strength of oscillatory alpha activity (Rihs, Michel, & Thut, 2007; Sauseng et al., 2005; Foxe, Simpson, & Ahlfors, 1998). An increase of alpha activity in the parieto-occipital area emerged when participants rested with their eyes closed (Salenius, Kajola, Thompson, Kosslyn, & Hari, 1995). Because the suppression of alpha activity was greater after the presentation of meaningful objects than after the presentation of meaningless nonobjects, it was thought that the difference of alpha activity was related to the strength of attention to the objects required by the task (Vanni, Revonsuo, & Hari, 1997). Ciesielski et al. (2007) showed that patients with obsessive compulsive disorder could maintain accurate performance despite an abnormally low alpha modulation in the parietooccipital area during the working memory task. Several studies have indicated that the strength of alpha activity is affected by the anticipatory attention effect (Haegens, Luther, & Jensen, 2012; Sauseng et al., 2005; Worden et al., 2000; Foxe et al., 1998). Our present results showed that alpha activity was suppressed not only by the presentation of distractors but also by the presentation of the visual cue, the end signal, the probe stimuli, and Japanese characters. We also found that, in some participants, alpha activity was suppressed before the distractor was presented. Thus, we can assume that alpha activity may be suppressed by the presentation or anticipation of the visual stimuli to which the participant needed to attend and an increase of alpha activity may reflect the disengagement of attention from the visual input. The present study showed that the increases of alpha activity occurred regardless of whether working memory was required. This suggests that alpha activity reflects disengagement rather than memory maintenance. The presentation of distractors, which had spatial information, reduced alpha activity. This can be explained by the notion that the presentation or anticipation of visual stimuli increased attention or reduced the disengagement. Thus, the present result supports the interpretation that the increase of alpha activity reflects the disengagement. In addition, although the suppression of alpha activity by the presentation of Japanese characters was transient, the suppression by the presentation of distractors was maintained. In this study, Japanese characters were presented at the center of the screen, whereas distractors were 462
Journal of Cognitive Neuroscience
presented around the characters and thus had spatial information. Alpha activity may be suppressed by spatial information. Therefore, we could conclude that the increase of alpha activity in the parieto-occipital area reflects the disengagement of attention from the visuospatial input. Acknowledgments The authors thank Dr. K. Takeda for offering valuable comments regarding the working memory task, Mr. A. Yoshida for providing technical support for the working memory task, and Ms. E. Takeda for providing support for MEG recording. This study was supported by a Grant-in-Aid for Young Scientists (B; 21700541, 24700542) to S. T. and a Grant-in-Aid for Scientific Research (C; 22500373) to T. N. from the Japan Society for the Promotion of Science. Reprint requests should be sent to Dr. Satoe Ichihara-Takeda, Department of Occupational Therapy, School of Health Science, Sapporo Medical University, South 1 West 17, Chuo-ku, Sapporo 060-8556, Japan, or via e-mail: [email protected].
REFERENCES Baddeley, A. (1986). Working memory. Oxford: Oxford University Press. Carlson, S., Martinkauppi, S., Rämä, P., Salli, E., Korvenoja, A., & Aronen, H. J. (1998). Distribution of cortical activation during visuospatial n-back tasks as revealed by functional magnetic resonance imaging. Cerebral Cortex, 8, 743–752. Chafee, M. V., & Goldman-Rakic, P. S. (1998). Matching patterns of activity in primate prefrontal area 8a and parietal area 7ip neurons during a spatial working memory task. Journal of Neurophysiology, 79, 2919–2940. Ciesielski, K. T., Hämäläinen, M. S., Geller, D. A., Wilhelm, S., Goldsmith, T. E., & Ahlfors, S. P. (2007). Dissociation between MEG alpha modulation and performance accuracy on visual working memory task in obsessive compulsive disorder. Human Brain Mapping, 28, 1401–1414. Curtis, C. E., & DʼEsposito, M. (2003). Persistent activity in the prefrontal cortex during working memory. Trends in Cognitive Science, 7, 415–423. Foxe, J. J., Simpson, G. V., & Ahlfors, S. P. (1998). Parieto-occipital approximately 10 Hz activity reflects anticipatory state of visual attention mechanisms. NeuroReport, 9, 3929–3933. Funahashi, S., Bruce, C. J., & Goldman-Rakic, P. S. (1989). Mnemonic coding of visual space in the monkeyʼs dorsolateral prefrontal cortex. Journal of Neurophysiology, 61, 331–349. Goldman-Rakic, P. S. (1987). Circuitry of primate prefrontal cortex and regulation of behavior by representational memory. In F. Plum (Ed.), Handbook of physiology, sect. 1, vol. V: The nervous system. Higher functions of the brain (pp. 373–417). Bethesda, MD: American Physiological Society. Haegens, S., Luther, L., & Jensen, O. (2012). Somatosensory anticipatory alpha activity increases to suppress distracting input. Journal of Cognitive Neuroscience, 24, 677–685. Ichihara-Takeda, S., & Funahashi, S. (2007). Activity of primate orbitofrontal and dorsolateral prefrontal neurons: Task-related activity during an oculomotor delayed-response task. Experimental Brain Research, 181, 409–425. Ichikawa, A., Yamamoto, H., Ono, I., Matsubayashi, J., Nagamine, T., Fukuyama, H., et al. (2007). Stimulus-related 20-Hz activity of human cortex modulated by the way of presenting hand actions. Neuroscience Research, 58, 285–290.
Volume 27, Number 3
Jensen, O., Gelfand, J., Kounios, J., & Lisman, J. E. (2002). Oscillations in the alpha band (9–12 Hz) increase with memory load during retention in a short-term memory task. Cerebral Cortex, 12, 877–882. Jensen, O., & Mazaheri, A. (2010). Shaping functional architecture by oscillatory alpha activity: Gating by inhibition. Frontiers in Human Neuroscience, 4, 186. Jokisch, D., & Jensen, O. (2007). Modulation of gamma and alpha activity during a working memory task engaging the dorsal or ventral stream. Journal of Neuroscience, 27, 3244–3251. Klimesch, W., Sauseng, P., & Hanslmayr, S. (2007). EEG alpha oscillations: The inhibition-timing hypothesis. Brain Research Reviews, 53, 63–88. Nagamine, T., Kajola, M., Salmelin, R., Shibasaki, H., & Hari, R. (1996). Movement-related slow cortical magnetic fields and changes of spontaneous MEG- and EEG-brain rhythms. Electroencephalography and Clinical Neurophysiology, 99, 274–286. Palva, J. M., Palva, S., & Kaila, K. (2005). Phase synchrony among neuronal oscillations in the human cortex. Journal of Neuroscience, 25, 3962–3972. Palva, S., & Palva, J. M. (2007). New vistas for alpha-frequency band oscillations. Trends in Neuroscience, 30, 150–158. Park, S., & Holzman, P. S. (1992). Schizophrenics show spatial working memory deficits. Archives of General Psychiatry, 49, 975–982. Park, S., Holzman, P. S., & Goldman-Rakic, P. S. (1995). Spatial working memory deficits in the relatives of schizophrenic patients. Archives of General Psychiatry, 52, 821–828. Pfurtscheller, G., Stancák, A., Jr., & Neuper, C. (1996). Event-related synchronization (ERS) in the alpha band—An electrophysiological correlate of cortical idling: A review. International Journal of Psychophysiology, 24, 39–46. Postle, B. R., & DʼEsposito, M. (1999). Dissociation of human caudate nucleus activity in spatial and nonspatial working memory: An event-related fMRI study. Brain Research: Cognitive Brain Research, 8, 107–115. Rihs, T. A., Michel, C. M., & Thut, G. (2007). Mechanisms of selective inhibition in visual spatial attention are indexed by alpha-band EEG synchronization. European Journal of Neuroscience, 25, 603–610. Roesch, M. R., & Olson, C. R. (2005). Neuronal activity dependent on anticipated and elapsed delay in macaque prefrontal cortex, frontal and supplementary eye fields, and premotor cortex. Journal of Neurophysiology, 94, 1469–1497. Sakai, K., Rowe, J. B., & Passingham, R. E. (2002). Active maintenance in prefrontal area 46 creates distractor-resistant memory. Nature Neuroscience, 5, 479–484.
Salenius, S., Kajola, M., Thompson, W. L., Kosslyn, S., & Hari, R. (1995). Reactivity of magnetic parieto-occipital alpha rhythm during visual imagery. Electroencephalography and Clinical Neurophysiology, 95, 453–462. Salmelin, R., Hämäläinen, M., Kajola, M., & Hari, R. (1995). Functional segregation of movement-related rhythmic activity in the human brain. Neuroimage, 2, 237–243. Salmelin, R., & Hari, R. (1994). Spatiotemporal characteristics of sensorimotor neuromagnetic rhythms related to thumb movement. Neuroscience, 60, 537–550. Sauseng, P., Klimesch, W., Heise, K. F., Gruber, W. R., Holz, E., Karim, A. A., et al. (2009). Brain oscillatory substrates of visual short-term memory capacity. Current Biology, 19, 1846–1852. Sauseng, P., Klimesch, W., Stadler, W., Schabus, M., Doppelmayr, M., Hanslmayr, S., et al. (2005). A shift of visual spatial attention is selectively associated with human EEG alpha activity. European Journal of Neuroscience, 22, 2917–2926. Schnitzler, A., Salenius, S., Salmelin, R., Jousmäki, V., & Hari, R. (1997). Involvement of primary motor cortex in motor imagery: A neuromagnetic study. Neuroimage, 6, 201–208. Tamura, Y., Hoshiyama, M., Nakata, H., Hiroe, N., Inui, K., Kaneoke, Y., et al. (2005). Functional relationship between human rolandic oscillations and motor cortical excitability: An MEG study. European Journal of Neuroscience, 21, 2555–2562. Thut, G., Nietzel, A., Brandt, S. A., & Pascual-Leone, A. (2006). Alpha-band electroencephalographic activity over occipital cortex indexes visuospatial attention bias and predicts visual target detection. Journal of Neuroscience, 26, 9494–9502. Tuladhar, A. M., ter Huurne, N., Schoffelen, J. M., Maris, E., Oostenveld, R., & Jensen, O. (2007). Parieto-occipital sources account for the increase in alpha activity with working memory load. Human Brain Mapping, 28, 785–792. Vanni, S., Revonsuo, A., & Hari, R. (1997). Modulation of the parieto-occipital alpha rhythm during object detection. Journal of Neuroscience, 17, 7141–7147. Watanabe, Y., & Funahashi, S. (2004). Neuronal activity throughout the primate mediodorsal nucleus of the thalamus during oculomotor delayed-responses: I. Cue-, delay-, and response-period activity. Journal of Neurophysiology, 92, 1738–1755. Worden, M. S., Foxe, J. J., Wang, N., & Simpson, G. V. (2000). Anticipatory biasing of visuospatial attention indexed by retinotopically specific alpha-band electroencephalography increases over occipital cortex. Journal of Neuroscience, 20, RC63.
Ichihara-Takeda et al.
463
Copyright of Journal of Cognitive Neuroscience is the property of MIT Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
