Brain, Vol. 122, No. 8, 1507-1518,
August 1999
© 1999 Oxford University Press
Orienting attention in time
Modulation of brain potentials
1 Department of Experimental Psychology, University of Oxford, Oxford and 2 Functional Imaging Laboratory, Wellcome Department of Cognitive Neurology, Institute of Neurology, London, UK
Correspondence to:
Anna C. Nobre, University of Oxford, Department of Experimental Psychology, South Parks Road, Oxford OX1 3UD, UK E-mail: anna.nobre{at}psy.ox.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
With the aim of casting light on the neural mechanisms that support our ability to modulate visual attention over time, we recorded event-related potentials (ERPs) while normal human subjects performed a target detection task with temporal contingencies between cue and target stimuli. The task used two central cues, which predicted (80% validity) when a subsequent target would occur (either 600 or 1400 ms after cue onset). Unlike previous tasks of attentional orienting, there was no spatial information provided and all stimuli were presented foveally. Reaction times and ERPs linked to targets presented at the shorter interval showed significant effects linked to attentional orienting. Reaction times were faster when the cues correctly predicted the cue–target interval, suggesting the ability of the brain to use information about time to deploy attentional resources. ERPs differed according to the predicted time interval. In particular, the P300 amplitude and latency were enhanced when the cue predicted the cue–target interval accurately. The ERPs elicited by the cues also differed according to the time interval that they predicted. Differences were observed in potentials linked to motor preparation and expectancies. The results reveal dynamic neural activity involved in orienting attention to time intervals, as well as the consequent modulation of target-related neural activity resulting from differing temporal expectations.
ANOVA = analysis of variance; CNV = contingent negative variation; ERP = event-related potential; SOA = stimulus onset asynchrony
temporal orienting; attention; ERPs; contingent negative variation; P300
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Our visual system is overloaded constantly with information from the environment, hence the advantage of efficient selective mechanisms for directing resources towards relevant stimuli. These selection mechanisms are one of the primary functions ascribed to visual attention. Since Posner described the behavioural effects of orienting attention to spatial locations (Posner, 1980
), there has been an enormous amount of experimentation on this issue. Cognitive psychological studies in humans suggest that we are able to focus attention dynamically on relevant locations in space, as revealed by improvements in the accuracy and speed of stimulus detection and discrimination. Studies using brain-damaged patients (Mesulam, 1990) or brain imaging of normal volunteers (Corbetta et al., 1993; Nobre et al., 1997) have revealed a frontoparietal network of brain regions that support spatial orienting of attention.
Experiments using event-related potentials (ERPs) have been especially useful in tracking the consequences that spatial expectations have upon subsequent stimulus processing. Brain activity engaged by stimuli at attended versus unattended locations is differentiated as early as the extrastriate visual cortex (Heinze et al., 1994; Woldorff et al., 1997). Visual-evoked potentials that reflect activity in extrastriate areas, namely P1 and N1, typically are enhanced in amplitude by attention, without observable changes in latency (see Mangun, 1995; Eimer, 1998).
Normal everyday experience would suggest that time intervals also play an important role in directing our attention to the external world. Our environment is dynamic and, consequently, time-related. The temporal dimension is an important aspect of perception and is necessary to guide effective action in the real world. If we are able to use information about where, why should we not be able to use information about when to expect an event? Despite the intuitive appeal, the ability to use temporal information to orient visual attention remains relatively unexplored.
Previous work on cognition relating to time has looked primarily at its subjective perception and estimation (Ruchkin et al., 1977; Macar and Vitton, 1980; Ladanyi and Dubrovsky, 1985). Our ability to use temporal information flexibly and actively to process a stimulus has not been investigated. Of course, any ability to use time information to direct attentional resources will depend on the mechanisms that support perception and estimation of temporal intervals. Brain regions involved in time perception and estimation have been revealed by neuropsychology (Nichelli et al., 1995, 1996; Harrington et al., 1998) and brain imaging (Jueptner et al., 1996). These regions include the prefrontal cortex, supplementary motor area, temporal lobe, basal ganglia and cerebellum (Goldman et al., 1968; Lacruz et al., 1991; Mangels et al., 1998).
The importance of time estimation processes for controlling behavioural performance has been addressed by experiments using warning signals preceding trials in cognitive tasks. Reaction times improve as a function of temporal certainty concerning the onset of stimuli requiring detection (Woodrow, 1914; Bertelson, 1967) or choice decisions (Bertelson and Boons, 1960; Broadbent and Gregory, 1965). Performance is improved when warning signals are delivered at constant, and therefore predictable, intervals. The processes that facilitate behaviour during the foreperiod anticipating the trial onset have been considered to be part of the alertness component of attention (Posner and Boies, 1971). The foreperiod is akin to a miniature situation of vigilance (Mackworth, 1970), during which alertness is developed rapidly and maintained briefly in preparation for behavioural performance.
In a recent study, we have demonstrated that information about time intervals can be used dynamically to direct visual attention (Coull and Nobre, 1998). In other words, the brief alertness effect observed in tasks with warning signals does not reflect a rigid process with a set optimal time course (e.g. Woodrow, 1914; Bertelson, 1967). Instead, deployment of anticipatory resources can be under flexible cognitive control. In the study by Coull and Nobre (Coull and Nobre, 1998), behavioural performance was improved by informative cues regarding the temporal intervals preceding the presentation of peripheral targets. Performance in the temporal orienting task engaged activity in frontoparietal networks, which overlapped extensively with brain areas engaged by spatial orienting tasks. However, within the common frontoparietal network, orienting attention in the two dimensions displayed an opposite pattern of hemispheric lateralization, with temporal orienting being relatively left-lateralized. Findings from patients with spatial attention deficits in the form of hemineglect also support a relationship between spatial attention and cognitive functions linked to processing of temporal intervals (Basso et al., 1996).
The present experiment used ERPs to explore the dynamics of the neural processes linked to directing attention toward temporal intervals, and to study the consequences of temporal orienting upon ongoing stimulus processing. The task differs from most previous studies of selective attention in that all stimuli were presented foveally. Thus, there was no spatial information that could be used for the selection or detection of targets (see also Valdes-Sosa et al., 1998; Nobre et al., 1999).
Analysis of ERPs elicited by the foveal target stimuli at predicted (attended) and unpredicted (unattended) time intervals should reveal how temporal orienting modulates ongoing brain activity. Comparison of the results with the consistent literature regarding modulation of visual ERPs by spatial orienting (see Mangun, 1995; Eimer, 1998; Müller et al., 1998) should help clarify the repertoire of attentional mechanisms available to influence behaviour. We were interested in observing whether temporal orienting also modulates early visual processing, in the same way as does the system for visual spatial orienting, or whether it acts at a different level of stimulus analysis, such as by influencing activity linked to motor functions.
Analysis of ERPs elicited by the cueing should reveal dynamics of brain activity linked to the orienting of attention in time. ERPs linked to cueing stimuli typically have not been reported (cf. Harter et al., 1989; Yamaguchi et al., 1994), perhaps because of the difficulty of equating the physical appearance of these stimuli across different cueing conditions. ERPs elicited by cueing stimuli should provide useful information about the on-line dynamics of the frontoparietal networks that direct attention, and how anticipatory activity interacts with target processing in sensorimotor areas.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Subjects
Twenty-one healthy right-handed adults (age range 18–38 years, 14 females) took part in the experiment as paid volunteers. Visual acuity was normal or corrected to normal. All participants reported being free of neurological disorders. The experimental methods were non-invasive and had ethical approval from the Department of Experimental Psychology, University of Oxford, UK.
Stimuli and task
The task, illustrated in Fig. 1, used two centrally placed cues that enabled subjects to predict when a subsequent target would occur. Each cue consisted of a narrow or wide cross (with an upper angle of 30 or 60°, respectively), and was flashed for 100 ms inside a circle of 1.7° of visual angle which was present continuously. The appearance of the narrow or wide cross cued the subject to expect either a short (600 ms) or a long (1400 ms) interval between cue onset and the appearance of the target. The designation of which cue corresponded to either a long or short interval was counterbalanced across subjects. The target consisted of the brightening of the circle surrounding the cue for 100 ms. The two possible intervals between cue and target occurred in an unpredictable sequence. The between trial onsets interval was randomized, and ranged between 3000 and 4300 ms.
|
Each subject performed 10 blocks of trials. The first two were practice blocks. Blocks consisted of 100 trials in total: 40 valid trials for each cue type (e.g. cue predicts long interval, target appears at 1400 ms); 10 invalid trials where the target occurred at the unpredicted time interval (e.g. cue predicts long interval, target appears at 600 ms); and 10 catch trials where no target occurred.
Procedure
Subjects were seated comfortably in a dimly illuminated, electrically shielded room, facing a computer monitor placed 100 cm in front of them. They were informed, at the beginning of the experiment, about the relationship between cue type and target. They maintained visual fixation on a small cross that was present continuously at the centre of the monitor. They were instructed to press a key with the index finger of the right or left hand as quickly as possible following onset of the target. Each subject used the left hand for half of the task and the right hand for the other half. The whole testing session lasted ~2 h, including short rest breaks after each block of trials. The subjects were instructed to suppress saccades and to avoid blinking during the task.
ERP recording
The EEG activity was recorded continuously from 54 sites using non-polarizable tin electrodes mounted on an elastic cap (Electro-Cap Inc.) and positioned according to the 10–20 International system (American Electroencephalographic Society, 1991). The montage included eight midline sites (FPZ, FZ, FCZ, CZ, CPZ, PZ, POZ and OZ) and 23 sites over each hemisphere (FP1/FP2, AF3/AF4, AF7/AF8, F3/F4, F5/F6, F7/F8, FC1/FC2, FC3/FC4, FC5/FC6, FT7/FT8, C3/C4, C5/C6, T7/T8, CP1/CP2, CP3/CP4, CP5/CP6, TP7/TP8, P3/P4, P5/P6, P7/P8, PO3/PO4, PO7/PO8 and O1/O2). Additional electrodes were used as ground and reference sites. The right mastoid served as the reference for all electrodes. Recordings obtained from a left mastoid electrode were used off-line to re-reference the scalp recordings to the average of the left and right mastoids. The signal was amplified 20 000 times and digitized at a sampling rate of 250 Hz. Data were recorded with a band-pass filter of 0.03–100 Hz.
The epoching of the ERPs was performed off-line. Epochs started 200 ms before and ended 1848 ms after stimulus onset. Separate ERP averaged waveforms were constructed from the four different types of targets (valid and invalid for each cue–target interval). ERPs elicited by the two types of cues (signalling short versus long intervals) were also constructed. ERPs to cues did not take into account trial validity, since this is not determined until the appearance of the target. These measures are, as a result, primarily a reflection of activity elicited during valid trials (80%), but also contain invalid (10%) and catch (10%) trials.
Horizontal and vertical eye movements were detected by recording the EOG. The voltage difference between two electrodes located lateral to the left and right external canthi recorded horizontal movements. The voltage difference between electrodes located above and beneath the right eye recorded vertical eye movements and blinks.
Epochs with eye movement artefacts (blinks or saccades) or incorrect behavioural responses were rejected. Cue-related epochs were discarded if the voltage exceeded ±100 µV between –200 and 1400 ms at any electrode site. In addition, epochs where voltage exceeded ±50 µV at the vertical and horizontal EOG electrodes between –200 and 600 ms were also discarded. The same voltage criteria were applied to target-related epochs between –200 and 600 ms.
Trials with reaction times faster than 130 ms or slower than 800 ms were regarded as errors, and were excluded. Data from subjects committing more than 25% errors on the catch trials were also eliminated from further consideration. To maintain an acceptable signal-to-noise ratio, a lower limit of 30 artefact-free trials per subject per condition was set.
ERP analysis
Identifiable visual ERP components (P1 and N1) were analysed for both cue and target stimuli. P1 and N1 were analysed at three lateral occipital electrode pairs: PO3/4, PO7/8 and O1/2. Measuring windows were determined from inspection of the group grand average waveforms. These latency windows were equivalent for ERPs elicited by both cues and targets. For each subject, the peak and latency of the P1 component were calculated at the time point of the largest positive peak between 60 and 140 ms. The peak and latency of the N1 component were calculated at the time point of the most negative peak between 100 and 200 ms. In addition, the P300 potential was analysed only for the target stimuli at electrode sites FZ, CZ and PZ. The P300 peak and latency were calculated at the largest positive-going peak occurring between 250 and 500 ms. All amplitude values were calculated with reference to the 200 ms pre-stimulus baseline for all components.
Data analysis was also performed using measures of the mean voltage value (mean amplitude) over successive time bins. We anticipated that these analyses would reveal the effects of task conditions on previously described ERP potentials that are linked to stimulus expectancy or motor preparation, as well as on ERP components for which we had no a priori hypotheses. Four region-specific sets of analyses were performed for cue stimuli. The exact electrodes used are listed during presentation of the results and are illustrated in the corresponding figures. Successive time bins in steps of 20 ms intervals between 200 and 600 ms were used for analyses over midline, frontal and posterior regions of the scalp. An additional analysis over the central scalp region used steps of 50 ms. Regional analyses for targets used mean amplitudes in 50 ms steps between 200 and 600 ms over midline and posterior sites.
Differences in latencies, peak amplitudes and mean amplitudes were assessed by repeated-measures analysis of variance (ANOVA), using the Greenhouse–Geisser epsilon correction for non-sphericity where appropriate (Jennings and Wood, 1976). The correction factor reduces the degrees of freedom of the usual F-test, and often results in non-integer values. Only the corrected probability values (and degrees of freedom) are reported.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Behavioural results
The behavioural data from 17 subjects were analysed. Four subjects were excluded: two committed >25% errors during the catch trials, and two had excessive artefacts during EEG recording.
Subjects contributing to the analysis showed high levels of accuracy in task performance (mean overall accuracy: 96.1%, range 92.1–99.4%). Reaction times to targets were subjected to an ANOVA that included cue assignment (wide/short, wide/long) as a between-subjects factor, and the following within-subjects factors: hand of response (left versus right), cue–target interval (short versus long) and target validity (valid versus invalid). The analysis revealed a main effect of validity [F(1,15) = 13.680; P = 0.0021], and an interaction between this factor and cue–target interval [F(1,15) = 28.433; P = 0.0001]. Figure 2 shows the pattern of findings. Advantages in reaction times as a function of trial validity occurred only at the short cue–target interval.
|
ERP results
Cues
Data from the same 17 subjects used for the behavioural analysis were used to characterize ERPs elicited by cues. Figure 3
displays ERP waveforms from the midline occipital (OZ) and central parietal (CPZ) sites, where the main ERP features are labelled. ERPs to cues were analysed according to its prediction (long or short interval).
|
Visual-evoked components
Cue stimuli elicited typical visual components P1 and N1 over posterior scalp sites (see Fig. 3
). There were no effects involving the factor of cue–target interval on measures of amplitude or latency. However, there were significant interactions between hemisphere and electrode, which characterized the distribution and time course of these potentials over the lateral occipital electrodes. The P1 had a larger amplitude over the right hemisphere, and was maximal at electrode PO8, as indicated by a hemisphere x electrode interaction [F(1.7,27.7) = 10.11; P < 0.0005]. The N1 potential had a maximal amplitude over posterior occipitotemporal sites (PO7/8), as shown by a main effect of electrode [F(1.4,22.7) = 15.578; P = 0.0001]. The latency also showed a significant main effect of electrode [F(1.1,17.3) = 10.68; P = 0.004], indicating that the fastest latency was recorded at PO3/4 sites.
Midline analysis
Following the N1, the ERPs were characterized by a positive deflection peaking at around 240 ms. Negative deflections followed, and were maximal over frontal and central sites, peaking at around 600 ms for the cue–target interval (see Fig. 3). Overall, the ERPs elicited by the short-interval cues were more negative than those elicited by long-interval cues following the initial visual responses.
An initial exploratory analysis was conducted on the following electrodes over the midline scalp: FPZ, FZ, CZ, PZ and OZ (Fig. 4). A within-subjects ANOVA tested the factors of cue–target interval (short interval versus long interval) and electrode (five sites). Effects involving the factor of cue–target interval were restricted to the 340–500 ms period. The main factor of cue–target interval was significant over the 360–380 and 400–420 ms time bins, and approached significance over the 380–400 ms period [360–380; F(1.0,16.0) = 5.197; P = 0.037; 380–400; F(1.0,16.0) =4.472; P = 0.051; 400–420; F(1.0,16.0) = 8.451; P =0.010]. These results reflect the fact that the ERPs to the short cue–target interval tend to be more negative than those to the long interval.
|
Interactions between interval and electrode site over the 340–380 ms time period reflected the fact that the differences between these ERPs were largest at the PZ electrode [340–360; F(2.4,38.7) = 4.823; P = 0.010; 360–380; F(2.3,36.1) =4.271; P = 0.018]. Over the 480–500 ms period, the cue–target interval x electrode interaction was again reliable [F(2.7,42.4) = 3.198; P = 0.038]. While the largest differences between these ERPs remained at PZ, in comparison with the earlier epochs, the magnitude of the differences at sites anterior to PZ increased, suggesting that the reasons for the interactions over the 340–380 and 480–500 ms periods were not equivalent. These apparent changes in the voltage distribution according to cue–target interval were assessed with directed analyses on normalized data (McCarthy and Wood, 1985
), as reported below.
Regional analyses
The frontal analysis included the following electrodes: FP1/2, AF3/4, AF7/8, F3/4, F5/6 and F7/8. Factors were cue–target interval (short versus long), hemisphere (left and right) and electrode (six sites). This analysis did not reveal any effects involving cue–target interval.
The posterior analysis included the same factors and the following electrodes: P3/4, P5/6, P7/8, PO3/4, PO7/8 and O1/2 (see Fig. 5). Main effects of cue–target interval over the 320–420 ms period reflected the fact that the ERPs to the short interval were relatively more negative than those to the long interval [for each 20 ms period F(1.0,16.0) >5.99; P < 0.05]. Three-way interactions involving cue–target interval, hemisphere and electrode were evident over the 280–300 and 300–320 ms epochs [280–300; F(3.1,49.5) =4.117; P = 0.010; 300–320; F(3.2,51.6) = 3.130; P =0.030]. These interactions reflected the fact that at the more anterior electrode pairings (P3/4 and P5/6) the relative negativity for the short cue–target interval is more pronounced over the left hemisphere, while at more posterior electrode pairings, in particular PO7/8, this hemisphere asymmetry is reversed. The only other effect involving cue–target interval was an interaction between this factor and electrode site from 540 to 560 ms [F(2.9,45.7) = 2.907; P = 0.047], reflecting the fact that the relative negativity for the short interval is evident principally at the P3/4 and PO3/4 electrode pairs.
|
The central analysis used 50 ms time bins between 200 and 600 ms, and included the following electrodes: FCZ, CZ, CPZ, FC3/4, C3/4 and CP3/4 (see Fig. 6
). Factors were cue–target interval (short versus long), hemisphere (left, midline and right) (the use of the term `hemisphere' is not strictly correct, but it avoids more confusing terminology) and electrode (three sites). The analyses revealed main effects of cue–target interval and interactions between cue–target interval and hemisphere at each 50 ms time bin between 350 and 500 ms (all Fs greater than 6.00, P < 0.05). These results indicated that the effects of cue–target interval are larger over the midline (FCZ, CZ and CPZ) than at the lateral electrode sites. There were no interactions with electrode location.
|
Directed topographical analyses
The effects linked to cue–target interval appeared to have different scalp distributions over time. The earliest effects were obtained at analyses restricted to the posterior region (280–300 ms). Over this early time period, no effects involving cue–target interval were evident for the midline analyses. Subsequent focal effects revealed by the midline analyses took the form of interactions between interval and electrode site (340–380 and 480–500 ms). As previously noted, the distributions of the differences between the long and short intervals did not appear to be equivalent over these later time periods. Considered jointly, these results suggest that the way in which the two cue–target intervals differ changes with time. This possibility was investigated using directed analyses on normalized data (McCarthy and Wood, 1985
). The effect of cue–target interval was isolated by subtracting waveforms elicited in the long cue–target interval condition from waveforms elicited in the short cue–target interval condition. Subtraction data from all electrode sites at the relevant time bins were re-scaled according to the minimum and maximum voltage values across all sites. Figure 7 shows colour-scaled topographies at the three time bins containing the reliable effects involving cue–target interval.
|
To compare the topographies of the early posterior effect (280–300 ms) with the first effects evident in the midline analysis (340–360 ms), a repeated-measures ANOVA was conducted over a sample of central and posterior electrode sites: CZ, PZ, C3/4 and P3/4. The factors tested were: time bin (280–300 versus 340–360 ms), hemisphere (left, midline and right) and electrode (two sites). An interaction between time bin and hemisphere showed that the lateral distribution of the topographies differed across the two bins [F(2.0,16.0) =3.29; P = 0.05]. The earlier effect was more laterally distributed and biased toward the left hemisphere, while the later effect was restricted to the midline sites (see Fig. 7
). To test the difference in the anterior–posterior extent of the effects obtained in the midline analysis, an ANOVA tested the factor of time bin (340–360 versus 480–500 ms) over the two electrode sites of interest (CZ versus PZ). A significant interaction between time bin and electrode confirmed that the topographies were distinct at the two bins [F(1.0,16.0) = 6.198; P = 0.024], with a relatively more posterior distribution at the earlier time bin.
In summary, analysis of ERPs elicited by the cue stimuli revealed the timing and the pattern of brain activity associated with focusing attention over time, in anticipation of upcoming targets. Changes in ERPs according to the expectation of long or short time intervals involved multiple brain processes and regions, as revealed by significant shifts in the scalp distribution of the effect.
Targets
In the analysis of target-related ERPs, only 10 of the 17 subjects reached the criteria for inclusion, due to insufficient trials in the target categories (see Methods). This smaller pool of subjects was still sufficient for reliably assessing possible modulations of target processing by selective orienting of time attention. Despite the smaller number of subjects, we believe these results are representative. They have been confirmed and extended by additional ongoing experiments in the laboratory (unpublished observations). Analysis of target validity was implemented separately for each cue–target interval. It was not possible to compare targets occurring at the two cue–target intervals directly, because a marked negative shift was present during the baseline activity for the targets at the short interval only (see Figs 3 and 8). The methodological problems regarding analysis of ERPs with different component overlap during the baseline period has been addressed previously in detail (Woldorff, 1993).
|
Visual-evoked components
There were no effects involving validity on amplitude or latency measures of P1 or N1 for targets at either stimulus onset asynchrony (SOA).
P300 component
The P300 was affected by target validity only for short SOA trials. The peak of the P300 was significantly earlier for expected validly cued targets (326 ms) than for targets that appeared unpredictably at the short interval (375 ms) [F(1.0,9.0) = 8.924; P = 0.015]. The peak amplitude of the P300 in short SOA trials was larger for expected targets than for invalidly cued targets (8.412 versus 6.984 µV), but this result did not reach statistical significance.
Midline and posterior analyses
Midline and posterior analyses of targets were performed over successive 50 ms bins from 200 to 600 ms. The midline analysis included sites FZ, CZ and PZ, and the posterior analysis included sites O1/2, PO7/8, PO3/4, P7/8, P5/6 and P3/4. Both sets of analyses revealed similar results. ERPs elicited by valid targets in the short-interval trials were more positive-going than ERPs elicited by invalid targets between 200 and 350 ms over the midline (Fs > 5.00, P < 0.05), and between 200 and 300 ms over the posterior region (Fs > 5.00, P < 0.05). These effects reflect modulation of the P300 component. For both midline and posterior analyses, there were no significant differences for ERPs elicited by targets at the long interval. Figures 8 and 9 show the ERPs elicited by targets at the short and long intervals over the midline and posterior regions, respectively.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The aim of the present experiment was to investigate the neural correlates of focusing visual attention selectively in time. The behavioural results of the study showed that we are indeed capable of using information about time intervals to improve our behavioural performance. An attentional effect was found with a faster reaction time for detection of valid versus invalid targets. However, reaction time effects were restricted to targets that occurred at the short interval. We suggest that this reflects the subjects' ability to re-focus their attention to the long interval once the expected target time has elapsed. These behavioural findings suggest that the brain is capable of using temporal information to deploy attentional resources dynamically. Temporal alertness effects (Woodrow, 1914
; Bertelson, 1967; Posner and Boies, 1971) are therefore under flexible conscious control according to predictable stimulus contingencies.
ERP recordings during task performance permitted on-line measures of shifts of attention, as well as the assessment of the consequences of temporal expectancy on visual and motor processing of the targets. The differences between cue types involved modulation of slow negative potentials with differing distributions over posterior and central scalp regions, suggesting the involvement of multiple brain areas and cognitive processes. Selective processing of the cue type started at around 280 ms over the posterior area, several hundred milliseconds before the target appeared. This initial modulation may reflect some aspect of the voluntary focusing of attention in time, which varies according to the symbolic meaning of the cue. Later differences between cue types occurred over the midline, with a distribution that shifted from parietal sites anteriorly with time (see Fig. 7). We hypothesize that these midline effects are modulations in processes that have been linked to the contingent negative variation (CNV) component (Walter et al., 1964; Weinberg, 1972). The CNV was first described by Walter (Walter et al., 1964) as a negative voltage change occurring between paired stimuli, when the first stimulus was a warning and the second one required a motor response or a decision. The CNV, originally described as an `expectancy' wave, more recently was divided in two components. One early component may be related to a non-motor preparatory process and one late component may be related to a motor preparatory process (Loveless and Sanford, 1974; Brunia, 1988). This early negative shift is a component present in tasks involving stimulus timing as described by Ruchkin (Ruchkin et al., 1977), and can be interpreted as an index of expectancy for an upcoming stimulus.
The neural generators of the effects described are difficult to define on the basis of the ERP findings alone. The scalp distribution would suggest an earlier sensitivity of posterior brain regions to the cueing information of the orienting stimulus, followed by activation of more anterior regions linked to motor preparation. Modulation of these posterior and central negativities is consistent with activation of parietal and premotor areas, observed in our previous brain imaging study of temporal orienting (Coull and Nobre, 1998). Furthermore, the ERP data also suggest that specific effects of temporal orienting have a left hemisphere bias (see topography at 280 ms, Fig. 7). However, the differences in the parameter details between the present ERP experiment and the previous brain imaging experiment (for instance, the brain imaging experiment presented peripheral target stimuli; Coull and Nobre, 1998) mean that correspondences between the two sets of data must be considered with caution. Further studies combining ERPs with brain imaging, or using source localization procedures, will be necessary to clarify the generators involved in the present task.
Modulation of target processing by temporal orienting was distinct in nature from previously described effects in visual spatial selective attention tasks (see Mangun, 1995). Temporal attention did not involve changes in amplitude or latency of the visual P1 and N1 in the present experiment. The observation contrasts with findings in previous spatial attentional studies using peripheral stimulation, where visual potentials of attended targets are relatively enhanced (Hillyard et al., 1973; Van Vooris and Hillyard, 1997; Mangun, 1995). It thus appears that selective temporal attention does not involve early perceptual modulation. However, we should not overlook the fact that all the stimuli were presented foveally, in contrast to previous studies. Visual processing of foveal stimuli is already optimized in the visual system. It might therefore not be necessary or advantageous to enhance resources further. Alternatively, even if perceptual selection processes were at work, the method may not have had the resolution or power to detect these.
Modulation of brain activity elicited by the target stimuli was evident after the visual components, and paralleled the pattern of behavioural results. The ERP modulation was restricted to targets that occurred at the short interval. This attentional effect had multiple aspects. Consistent with many reports in studies of spatial attention (reviewed by Näätänen et al., 1992; Hillyard et al., 1996), P300 amplitude was larger for valid targets (though the effect reached statistical reliability only with the area measures). The significant modulation of P300 latency was also striking. The P300 peak was earlier for valid targets. Latency modulation of the P300, or other components, typically has not been reported during spatial attention tasks. The finding therefore suggests that temporal orienting may affect stimulus processing by mechanisms different from or in addition to those previously observed during spatial attention ERP tasks. Information about temporal intervals may be used to synchronize and/or prepare motor processes, or sharpen processes linked to decisions or responses.
Latency modulation of the P300 has been found in tasks that manipulate response requirements to stimuli. Stimuli signalling the requirement for a response (Go signals) elicit a P300 which peaks earlier and with a more posterior distribution than the P300 elicited by stimuli signalling the withholding of a response (No-go signals) (e.g. Hillyard et al., 1996; Simson et al., 1977). In Go/No-go tasks, the latency modulation of the P300 could reflect different response- or decision-related processing of the two types of signal. Alternatively, it could be determined by differences in the resolution times of preceding CNV potentials in the two cases (see Eimer, 1993). Similarly, the present task may manipulate aspects of motor preparation to the target stimuli. After a cue that signals a long interval, subjects may actively withhold the response (akin to the No-go situation) until the appropriate later time. Overlap with potentials linked to anticipation or motor preparation could also have influenced the P300 latency in the present study.
Modulation of the N2 potential has also been observed in Go/No-go tasks. These potentials are enhanced to No-go stimuli, and have been suggested to reflect the detection of a deviation from the established stimulus–response associations (Mantysälo, 1987) or a response inhibition process (Kok, 1986). Response inhibition interpretations have been favoured by results of tasks which have manipulated Go:No-go probabilities and the orienting of spatial attention (Eimer, 1993). In the present experiment, invalid trials elicited larger posterior negativities between 200 and 300 ms (this contrast was reliable only for the targets at the short interval, though parallel effects are observable in the waveforms elicited by targets at the long interval (see Fig. 9). However, it is unlikely that these effects are the same as those observed in previous Go/No-go attention experiments (see Eimer, 1993). The distribution of the negative potential was posterior, and the potentials were not always enhanced when responses had to be withheld. In the case of the targets appearing at the short interval, the situation was the reverse. The negative potentials were related to the releasing of a response. More generally, the negative potentials were evoked in conditions with breaches in the expected stimulus associations that guide responses (see Nobre et al., 1999). The exact functional significance of this effect must await further experimentation.
In keeping with most ERP studies of attention, our analysis emphasized the comparison of transient activity engaged by the target stimuli, such as the visual-evoked responses. It eliminated possible contributing effects from differences in the level or distribution of brain activity directly preceding target onset (i.e. we normalized the baseline activity at each site to a 0 µV level). However, analysis during the cue–target interval indicated that more tonic changes in the activation level preceding the target also contributed to the selective attention effect (e.g. for a similar issue in single-unit recording studies, see Luck et al., 1997). It is possible, and indeed likely, that real changes in baseline levels of activation have been present in most attentional cueing tasks to date, including many of the spatial attention tasks, though these effects typically have not been examined.
In conclusion, we have demonstrated that selective information about temporal intervals can be used dynamically to enhance behavioural performance during a target detection task, even in the absence of any spatial information. ERPs linked to shifts of attention in the temporal domain started several hundred milliseconds prior to the appearance of the subsequent target stimuli. Task effects on ERPs elicited by orienting cues suggested differential engagement of motor preparation for short versus long time intervals. Modulation of target–stimulus processing by selective temporal attention was also more related to changes in later potentials, which may reflect changes in response preparation or decision. Modulation of early visual potentials linked to perceptual factors was not evident. The pattern of modulation of ongoing stimulus (target) processing is therefore quite distinct from that consistently observed during spatial orienting tasks. Selective attention can therefore modify different levels of stimulus analysis (perceptual versus response-related) according to whether spatial or temporal information is available.
|
Acknowledgments |
|---|
We wish to thank Leonardo Chelazzi for helpful comments on the manuscript, and Anling Rao for technical assistance. The research was supported by a project grant to A.C.N. from The Wellcome Trust.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
American Electroencephalographic Society. American Electroencephalographic Society guidelines for standard electrode position nomenclature [see comments]. J Clin Neurophysiol 1991; 8: 200–2. Comment in: J Clin Neurophysiol 1993; 10: 123–8.[Web of Science][Medline]
Basso G, Nichelli P, Frassinetti F, di Pellegrino G. Time perception in a neglected space. Neuroreport 1996; 7: 2111–4.[Web of Science][Medline]
Bertelson P. The time course of preparation. Q J Exp Psychol 1967; 19: 272–9.[Web of Science][Medline]
Bertelson P, Boons JP. Time uncertainty and choice reaction time. Nature 1960; 187: 531–2.[Medline]
Broadbent DE, Gregory M. On the interaction of S–R compatibility with other variables affecting reaction time. Br J Psychol 1965; 56: 61–7.
Brunia CH. Movement and stimulus preceding negativity. [Review]. Biol Psychol 1988; 26: 165–78.[Web of Science][Medline]
Corbetta M, Miezin FM, Shulman GL, Petersen SE. A PET study of visuospatial attention. J Neurosci 1993; 13: 1202–26.[Abstract]
Coull JT, Nobre AC. Where and when to pay attention: the neural systems for directing attention to spatial locations and to time intervals as revealed by both PET and fMRI. J Neurosci 1998; 18: 7426–35.
Eimer M. Effects of attention and stimulus probability on ERPs in a Go/Nogo task. Biol Psychol 1993; 35: 123–38.[Web of Science][Medline]
Eimer M. Mechanisms of visuospatial attention: evidence from event-related potentials. Vis Cogn 1998; 5: 257–86.
Goldman PS, Lodge A, Hammer LR, Semmes J, Mishkin M. Critical flicker frequency after unilateral temporal lobectomy in man. Neuropsychologia 1968; 6: 355–63.[Web of Science]
Harrington DL, Haaland KY, Knight RT. Cortical networks underlying mechanisms of time perception. J Neurosci 1998; 18: 1085–95.
Harter MR, Miller SL, Price NJ, LaLonde ME, Keyes AL. Neural processes involved in directing attention. J Cogn Neurosci 1989; 1: 223–37.
Heinze HJ, Mangun GR, Burchert W, Hinrichs H, Scholz M, Munte TF, et al. Combined spatial and temporal imaging of brain activity during visual selective attention in humans. Nature 1994; 372: 543–6.[Medline]
Hillyard SA, Hink RF, Schwent VL, Picton TW. Electrical signs of selective attention in the human brain. Science 1973; 182: 177–80.
Hillyard SA, Anllo-Vento L, Clark VP, Heinze HJ, Luck SJ, Mangun GR. Neuroimaging approaches to the study of visual attention: a tutorial. In: Kramer AF, Coles MG, Logan GD, editors. Converging operations in the study of visual selective attention. Washington (DC): American Psychological Association; 1996. p. 107–38.
Jennings JR, Wood CC. The epsilon-adjustment procedure for repeated-measures analyses of variance [letter]. Psychophysiology 1976; 13: 277–8.[Web of Science][Medline]
Jueptner M, Flerich L, Weiller C, Mueller SP, Diener H-C. The human cerebellum and temporal information processing—results from a PET experiment. Neuroreport 1996; 7: 2761–5.[Web of Science][Medline]
Kok A. Effects of degradation of visual stimulation on components of the event-related potential (ERP) in Go/Nogo reaction tasks. Biol Psychol 1986; 23: 21–38.[Web of Science][Medline]
Lacruz F, Artieda J, Pastor MA, Obeso JA. The anatomical basis of somaesthetic temporal discrimination in humans. J Neurol Neurosurg Psychiatry 1991; 54: 1077–81.
Ladanyi M, Dubrovsky B. CNV and time estimation. Int J Neurosci 1985; 26: 253–7.[Web of Science][Medline]
Loveless NE, Sanford AJ. Slow potential correlates of preparatory set. Biol Psychol 1974; 1: 303–14.[Medline]
Luck SJ, Chelazzi L, Hillyard SA, Desimone R. Neural mechanisms of spatial selective attention in areas V1, V2, and V4 of macaque visual cortex. J Neurophysiol 1997; 77: 24–42.
Macar F, Vitton N. CNV and reaction time task in man: effects of inter-stimulus interval contingencies. Neuropsychologia 1980; 18: 585–90.[Web of Science][Medline]
Mackworth JF. Vigilance and attention: a signal detection approach. Harmondsworth (UK): Penguin; 1970.
Mangels JA, Ivry RB, Shimizu N. Dissociable contributions of the prefrontal and neocerebellar cortex to time perception. Brain Res Cogn Brain Res 1998; 7: 15–39.[Medline]
Mangun GR. Neural mechanisms of visual selective attention. [Review]. Psychophysiology 1995; 32: 4–18.[Web of Science][Medline]
Mantysälo S. N2 and P3 of the ERP to Go and Nogo stimuli: a stimulus–response association and dissociation. Electroencephalogr Clin Neurophysiol Suppl 1987; 40: 227–34.[Medline]
McCarthy G, Wood CC. Scalp distributions of event-related potentials: an ambiguity associated with analysis of variance models. Electroencephalogr Clin Neurophysiol 1985; 62: 203–8.[Web of Science][Medline]
Mesulam MM. Large-scale neurocognitive networks and distributed processing for attention, language, and memory. [Review]. Ann Neurol 1990; 28: 597–613.[Web of Science][Medline]
Müller MM, Teder-Sälejärvi W, Hillyard SA. The time course of cortical facilitation during cued shifts of spatial attention. Nature Neurosci 1998; 1: 631–4.[Web of Science][Medline]
Näätänen R, Teder W, Alho K, Lavikainen J. Auditory attention and selective input modulation: a topographical ERP study. Neuroreport 1992; 3: 493–6.[Web of Science][Medline]
Nichelli P, Clark K, Hollnagel C, Grafman J. Duration processing after frontal lobe lesions. Ann NY Acad Sci 1995; 769: 183–90.[Web of Science][Medline]
Nichelli P, Alway D, Grafman J. Perceptual timing in cerebellar degeneration. Neuropsychologia 1996; 34: 863–71.[Web of Science][Medline]
Nobre AC, Sebestyen GN, Gitelman DR, Mesulam MM, Frackowiak RS, Frith CD. Functional localization of the system for visuospatial attention using positron emission tomography. Brain 1997; 120: 515–33.
Nobre AC, Coull JT, Frith CD, Mesulam MM. Orbitofrontal cortex is activated during breaches of expectation in tasks of visual attention. Nature Neurosci 1999; 2: 11–2.[Web of Science][Medline]
Pashler HE. The psychology of attention. Cambridge (MA): MIT Press; 1998.
Posner MI. Orienting of attention. Q J Exp Psychol 1980; 32: 3–25.[Web of Science][Medline]
Posner MI, Boies SJ. Components of attention. Psychol Rev 1971; 78: 391–408.[Web of Science]
Ruchkin DS, McCalley MG, Glaser EM. Event related potentials and time estimation. Psychophysiology 1977; 14: 451–5.[Web of Science][Medline]
Simson R, Vaughan HG Jr, Ritter W. The scalp topography of potentials in auditory and visual Go/NoGo tasks. Electroencephalogr Clin Neurophysiol 1977; 43: 864–75.[Web of Science][Medline]
Valdes-Sosa M, Bobes MA, Rodriguez V, Pinilla T. Switching attention without shifting the spotlight: object-based attentional modulation of brain potentials. J Cogn Neurosci 1998; 10: 137–51.[Web of Science][Medline]
Van Voorhis S, Hillyard SA. Visual evoked potentials and selective attention to points in space. Percept Psychophys 1977; 22: 54–62.[Web of Science]
Walter WG, Cooper R, Aldridge VJ, McCallum WC, Winter AL. Contingent negative variation: an electrical sign of sensorimotor association and expectancy in the human brain. Nature 1964; 203: 380–4.[Medline]
Weinberg H. The contingent negative variation: its relation to feedback and expectant attention. Neuropsychologia 1972; 10: 299–306.[Web of Science][Medline]
Woldorff MG. Distortion of ERP averages due to overlap from temporally adjacent ERPs: analysis and correction. Psychophysiology 1993; 30: 98–119.[Web of Science][Medline]
Woldorff MG, Fox PT, Matzke M, Laucaster JL, Veeraswamy S, Zamerripa F, et al. Retinotopic organization of early visual spatial attention effects as revealed by PET and ERPs. Hum Brain Mapp 1997; 5: 280–6.[Web of Science]
Woodrow H. The measurement of attention. Psychol Monographs 1914; 17 (No. 76): 1–158.
Yamaguchi S, Tsuchiya H, Kobayashi S. Electroencephalographic activity associated with shifts of visuospatial attention. Brain 1994; 117: 553–62.
Received March 7, 1999. Accepted March 18, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Bueti and V. Walsh The parietal cortex and the representation of time, space, number and other magnitudes Phil Trans R Soc B, July 12, 2009; 364(1525): 1831 - 1840. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. G. O'Connell, P. M. Dockree, I. H. Robertson, M. A. Bellgrove, J. J. Foxe, and S. P. Kelly Uncovering the Neural Signature of Lapsing Attention: Electrophysiological Signals Predict Errors up to 20 s before They Occur J. Neurosci., July 1, 2009; 29(26): 8604 - 8611. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. M. BAUSENHART, B. ROLKE, and R. ULRICH Temporal preparation improves temporal resolution: Evidence from constant foreperiods Atten Percept Psychophys, November 1, 2008; 70(8): 1504 - 1514. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Correa and A. C. Nobre Neural Modulation by Regularity and Passage of Time J Neurophysiol, September 1, 2008; 100(3): 1649 - 1655. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. D. SANDERS and L. B. ASTHEIMER Temporally selective attention modulates early perceptual processing: Event-related potential evidence Atten Percept Psychophys, May 1, 2008; 70(4): 732 - 742. [Abstract] [PDF]
|
|
|
|
 |
|
 P. Lakatos, G. Karmos, A. D. Mehta, I. Ulbert, and C. E. Schroeder Entrainment of Neuronal Oscillations as a Mechanism of Attentional Selection Science, April 4, 2008; 320(5872): 110 - 113. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Fan, J. Byrne, M. S. Worden, K. G. Guise, B. D. McCandliss, J. Fossella, and M. I. Posner The Relation of Brain Oscillations to Attentional Networks J. Neurosci., June 6, 2007; 27(23): 6197 - 6206. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Praamstra, D. Kourtis, H. Fei Kwok, and R. Oostenveld Neurophysiology of implicit timing in serial choice reaction-time performance. J. Neurosci., May 17, 2006; 26(20): 5448 - 5455. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Eagleman, P. U. Tse, D. Buonomano, P. Janssen, A. C. Nobre, and A. O. Holcombe Time and the Brain: How Subjective Time Relates to Neural Time J. Neurosci., November 9, 2005; 25(45): 10369 - 10371. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. R. Doherty, A. Rao, M. M. Mesulam, and A. C. Nobre Synergistic Effect of Combined Temporal and Spatial Expectations on Visual Attention J. Neurosci., September 7, 2005; 25(36): 8259 - 8266. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Seiss and P. Praamstra The basal ganglia and inhibitory mechanisms in response selection: evidence from subliminal priming of motor responses in Parkinson's disease Brain, February 1, 2004; 127(2): 330 - 339. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.T. Coull, A.C. Nobre, and C.D. Frith The Noradrenergic {{alpha}}2 Agonist Clonidine Modulates Behavioural and Neuroanatomical Correlates of Human Attentional Orienting and Alerting Cereb Cortex, January 1, 2001; 11(1): 73 - 84. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||