- Email: [email protected]
Numerical magnitude modulates temporal comparison: An ERP study
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –1 42
a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m
w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s
Research Report
Numerical magnitude modulates temporal comparison: An ERP study Bin Xuan a,b , Xiang-Chuan Chen a , Sheng He c , Da-Ren Zhang a,⁎ a
Division of Bio-X Interdisciplinary Sciences, Hefei National Laboratory for Physical Sciences at Microscale and Department of Neurobiology and Biophysics, School of Life Science, University of Science and Technology of China, Hefei, Anhui 230027, PR China b Department of Psychology, Anhui Normal University, Wuhu, Anhui 241000, PR China c Department of Psychology, University of Minnesota, 75 East River Road, Minneapolis, MN 55455, USA
A R T I C LE I N FO
AB S T R A C T
Article history:
Time is believed to be a part of the generalized magnitude system just like space and
Accepted 6 March 2009
quantity. Previous research suggests that time perception can be affected by magnitude in
Available online 21 March 2009
some non-temporal dimensions. Here we address two questions. First, could the influence be caused by an abstract magnitude component without perceptual variables? Second, what
Keywords:
are the underlying mechanisms of the influence? Participants compared a pair of durations
Numeric magnitude
defined by two Arabic digits in a hundreds of milliseconds range. They performed more
Temporal comparison
accurately when the shorter durations were defined by lower numeric value digits (small
Selective temporal attention
digits) and the longer durations were defined by higher value digits (large digits) than they
Contingent negative variation
did in the reversed condition. Event-Related Potential (ERP) results showed that the CNVs
N1
corresponding to the first duration (CNV1), to the second duration (CNV2) and the N1 were
Event-Related Potential
all enhanced when durations marked by small digits than that marked by large ones. Combining the electrophysiological data with the behavioral results, we suggest that digits can modulate performance of temporal comparison at the relatively early stage of perceptual processing. One possible explanation of the current results is that selective temporal attention and subsequent expectation may be involved in this modulation. © 2009 Elsevier B.V. All rights reserved.
1.
Introduction
Time is a fundamental dimension of our existence. Processing of temporal information is crucial to many aspects of our daily lives, from our sleep–wake cycle to speech recognition, music perception and skilled motor control, even just waiting an upcoming event (Buhusi and Meck, 2005). We can experience time across a wide range of intervals, but often represent and estimate it subjectively with limited precision, especially in a hundreds of milliseconds range, for its susceptibility to many environmental factors and cognitive processes.
Time is traditionally believed to be a fundamentally different perceptual dimension from space or quantity. However, increasing evidence indicates that perceiving temporal magnitude may not be an independent or specialized neural process, but has close relationships with perceiving other magnitude dimensions in the generalized magnitude system (Walsh, 2003). Take “the Tau effect” as an example, when three points on observer's forearm were stimulated in succession to define two spatial and temporal information, Helson found that the judgments of relative positions were high correlated with the relative intervals. If the temporal interval between
⁎ Corresponding author. Fax: +86 551 3601443. E-mail address: [email protected] (D.R. Zhang). 0006-8993/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2009.03.016
136
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –14 2
stimulating the second and third points was greater than that between the first and second, observers were inclined to report that the tactile spatial distance between the second and third points was greater than that between the first and second, though in fact it might be equal or less (Helson, 1930). Similarly, the judgments of temporal intervals between two stimuli were also reported to depend on their spatial relations (Cohen et al., 1953). It was even observed that the experience of time is compressed together with space in scale-model environments (DeLong, 1981), which reflects the psychologically interdependence between magnitude information in spatial and temporal dimensions. Electrophysiological, neuropsychological and further behavioral studies all implicate the possible spatialrelated timing mechanisms (Basso et al., 1996; Buonomano and Merzenich, 1995; Grondin, 1998; Johnston et al., 2006). On the other hand, the relationships between perceiving time and magnitude in other dimensions were also observed. Dormal et al. found that judging the duration of dot sequences is influenced by the irrelevant number of dots (Dormal et al., 2006). Our previous study also suggest the error rates of temporal judgment could be significantly affected by the magnitudes in non-temporal dimensions, including number of dots, size of open squares, luminance of solid squares, and numeric value of digits (Xuan et al., 2007). In addition, quite a few research also demonstrated that number, another part of the proposed generalized magnitude system, and space are also close related. Dehaene et al. made the pioneer studies of this domain. They put forward “number sense” — a domainspecific, biologically-determined ability (Dehaene et al., 1998), but they also indicated that numbers can be represented and manipulated on a “mental number line” (Dehaene et al., 1993). This analogical spatial representation robustly influences the performance of many spatial cognitive tasks, such as magnitude comparison (Pinel et al., 2004), target detection (Fischer et al., 2003), line and number bisection (Doricchi et al., 2005; Fischer, 2001) etc. Based on the mentioned evidence, Walsh proposed a common magnitude concept. The magnitude concept is rooted in people's need for information about the nature of the external world, and it is often represented as “how many, how much, how long, how far and how fast” with shared metric, processing resources and behavioral goals (Walsh, 2003). However, there still remain disagreements on the relationship between time and magnitude in other dimensions until recently (Lewis and Walsh, 2005). First, some research indicates that short time on the scale of tens to hundreds of milliseconds may not be encoded explicitly as a metric (Burr and Morrone, 2006). Second, it is still a question whether there is an abstract and generalized magnitude component influence time perception. Specifically, the magnitude component is irrelevant to any perceptual variable, such as complexity, luminance, velocity of imputed motion or spatial attention shift (Cohen et al., 1953; Jones and Huang, 1982; Schiffman and Bobko, 1974; Schiffman and Bobko, 1977). To avoid these possible interferential factors, in the present study, Arabic digits were selected to mark durations in a temporal comparison task, and the control stimuli were also used to match any confounding physical parameter. Our first concern focused on whether the subjective time perception is affected by
abstract symbols bearing magnitude meanings such as Arabic digits, and whether the influence is independent of perceptual variables. If digits did modulate the subjective time perception, the underlying mechanisms of the modulation naturally became our next concern. Lots of studies have indicated that time perception is intimately related to the level of attention (Grondin, 2001; Macar et al., 1994; Perbal et al., 2003; Pouthas and Perbal, 2004; Sevigny et al., 2003). Both the accuracy (Coull et al., 2004) and the length (Tse et al., 2004) of subjective time perception can be influenced by attention resource deployment. What is more, Nobre, Coull and their colleagues simultaneously explored two fundamental cognitive functions: selective attention and processing of time. They gave a great impetus to the study of orienting attention to time. Similar to attention in spatial domain, a series of delicate experiments demonstrate that attention in the temporal domain can be directed to certain time point of an upcoming event intentionally, which is named as selective temporal attention (Coull and Nobre, 1998). With a modified Posner's paradigm (Posner, 1980), it was found that subjects can make use of valid temporal cues to correctly predict specific time points in high proportion, effectively improve the performance of target detection and accuracy of predictable temporal information (Correa et al., 2006b; Nobre, 2001). In the mentioned studies, cues were originally meaningless. The meanings of cues were set up through the proportion of correct predictions. However, digits themselves can convey meanings of magnitude. Fischer et al. investigated the use of digits as cues for the detection of lateralized spatial targets. They found that digits can direct spatial attention to the left or right of visual space (Fischer et al., 2003). Similarly, we suspected that digits might play a role of “temporal cue” in the present experiment. That is, digits may direct subjects to deploy attention to a certain time point according to their numerical magnitude. Small digits orient attention to early points in time, and large digits orient attention to late points in time. Based on this hypothesis, if small digits marked shorter duration, large digits marked longer duration, then subjects would establish valid time attention to expect the target — the ending of duration. That is, subjects would expect target arrived earlier with small digits and later with large digits, and it was consistent with this kind of experimental condition; In contrast, if large digits marked shorter duration and small digits marked longer duration, time attention would be invalid. That is, subjects generate a reversed expectation induced by digits as compared with this kind of experimental condition. The valid or invalid time attention would influence the accuracy of temporal comparison. To further validate the hypothesis, we use Event-Related Potential (ERP) to monitor the brain activity. ERP investigation is appropriate to track on-line the dynamic modulation processing of time perception for the high temporal resolution (Macar and Vidal, 2004). If temporal attention did mediate the influence, we can expect that the related ERP components, such as CNV were modulated by digits, since some previous studies indicated that CNV component was related with time attention and the expectation of the upcoming event (Griffin et al., 2002; Miniussi et al., 1999). The combined behavioral performance and the related
137
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –1 42
Fig. 1 – (A) Arabic digits and the control figures (scrambled images of digits) are used to define durations in the experiment. “Small” and “large” were assigned based on digit value. (B) The response accuracy of temporal judgment in the digit (left) and control figure (right) conditions. Subjects tended to judge large digits to last longer (*p < 0.006), which did not happen in control figures. Error bars represent one standard error.
brain potentials would cast light on the influence by digits in a temporal comparison task.
2.
Results
We used two types of stimuli to define duration: Arabic digits and control figures (scrambled digits). See Fig. 1A. There were two durations (one shorter and one longer) defined by two stimuli (one “large” and one “small”) in each trail. Participants were required to compare the lengths of two durations in each trial.
2.1.
Behavioral results
The behavioral performances of 14 subjects, including response accuracy and reaction time were analyzed by two-way repeated measures ANOVA with magnitude (small and large) and duration length (shorter and longer) according to the first stimuli configurations of each trial (see Table 1). There is no duration length main effect (digits: F(1,13)= 1.807, p > 0.202; control figures: F(1,13)= 1.480, p > 0.245) and no magnitude main effect (digits: F
(1,13) = 2.493, p > 0.138; control figures: F(1,13) = 0.630, p > 0.441) in response accuracy. However, there is a significant interaction between duration length and magnitude for digit condition (F (1,13) = 10.843, p < 0.006). That is, the accuracy of temporal comparison in the condition of shorter durations defined by small digits and longer durations defined by large digits is higher than in the reversed condition (see Fig. 1B). We divided the first digit of each trial into eight conditions (1, 2, 3, 4, 6, 7, 8 and 9) in the further detailed analysis. The accuracy of temporal comparison showed a significant magnitude main effect. When the first duration was shorter, accuracy tended to decrease with the increase of numeric value, F(7,91)= 2.884, p < 0.010; when the first duration was longer, accuracy tended to increase with the increase of numeric value, F(7,91) = 5.185, p < 0.0001. However, no similar interaction was found for control condition (F(1,13) = 0.091, p > 0.767). The statistic results of reaction time showed no duration length main effect (digits: F(1,13) = 4.335, p > 0.057; except for control condition: F(1,13) = 20.695, p < 0.001), no magnitude main effect (digits: F(1,13) = 0.567, p > 0.464; control figures: F(1,13)= 0.845, p > 0.374), and no interaction between the two factors (digit: F(1,13)= 1.251, p > 0.283; control figures: F(1,13)= 0.739, p > 0.405).
2.2.
ERP results
2.2.1.
Mean amplitudes of CNVs
Corresponding to the two durations in each trial, the two CNV components were elicited on 400 ms or so after each duration onset. The representative CNV1 (corresponding to the first duration) and CNV2 (corresponding to the second duration) are shown in Fig. 2A (for example, the CNVs elicited by 700–800 duration are shown). According to the previous studies (Correa et al., 2006a; Macar and Vidal, 2004) and our results, electrode grouping was performed in order to define a region of interest (ROI). The ROI includes 9 electrodes of frontal–parietal central site: F1, FZ, F2, FC1, FCZ, FC2, C1, CZ, and C2. The mean amplitudes of CNVs were measured over the four types of temporal window in accordance to the four types of durations. The temporal window ranged from 400 ms to the end of the duration (400–500 ms for duration of 500 ms; 400–600 ms for 600 ms; 400–700 ms for 700 ms; 400–800 ms for 800 ms, see Fig. 2B). Identical analysis was performed on both CNV1 and CNV2.
Table 1 – Response accuracy and reaction time. Digits Accuracy (%) Small Short Long dur (F,p) mag (F,p) dur × mag (F,p)
Large
74.26 62.05 (16.07) (19.75) 66.82 84.48 (17.80) (10.20) 1.807, 0.202 2.493, 0.138 10.843, 0.006
Control figures Reaction time (ms) Small
Large
537.03 559.31 (172.04) (165.66) 608.92 602.55 (151.58) (174.48) 4.335, 0.058 0.567, 0.465 1.251, 0.284
Accuracy (%) Small
Large
67.48 66.74 (15.06) (18.73) 71.35 73.66 (14.25) (14.13) 1.480, 0.245 0.630, 0.441 0.091, 0.768
Reaction time (ms) Small
Large
526.67 527.91 (177.65) (178.19) 622.58 602.62 (158.71) (166.86) 20.695, 0.001 0.845, 0.375 0.739, 0.406
Data in parentheses denote one standard error for all of the four conditions (represented with the first stimuli configuration of each trial: duration1/digit1). Here “dur” means duration length; “mag” means numeric magnitude.
138
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –14 2
Fig. 2 – (A) The red line indicates the representative CNVs at FCZ (including the CNV1 and the CNV2). CNV results were analyzed from the 9 electrodes of the red square. The blue line indicates the representative P1 and N1 components at PO5. The statistics results of P1s and N1s came from the 4 electrodes in the two blue diamonds. (B) The mean amplitudes of CNVs were measured over the temporal window according to the four types of duration: 400–500 ms was used for duration of 500 ms, 400–600 ms for 600 ms, 400–700 ms for 700 ms, and 400–800 ms for 800 ms.
The CNV results were analyzed by three-way repeated measures ANOVA with magnitude (small and large), temporal length (500, 600, 700 and 800 ms) and electrodes (9 electrodes). All of the results were subject to the Greenhouse–Geisser epsilon correction. The statistic results of the mean amplitudes of the CNV1 and the CNV2 show that the magnitude has main effect in digit condition (the CNV1: F(1,13) = 8.723, p < 0.012; the CNV2: F(1,13) = 5.332, p < 0.038). It is observed that the two CNV components elicited by small digits are both enhanced than those by large digits (Fig. 3A). Scalp distribution map shows that the difference in the frontal–parietal area already emerges at the initial stage of CNVs (from 400 to 500 ms) for all of the duration length conditions (Fig. 3B). No main effect of temporal length is found (the CNV1: F(3,39)=1.290, p > 0.289; the CNV2: F(3,39) = 0.304, p > 0.750), and no magnitude × temporal length interaction is found (the CNV1: F(3,39) = 0.127, p > 0.923; the CNV2: F(3,39) = 0.383, p > 0.728). No main effect of magnitude (the CNV1: F(1,13) = 1.204, p > 0.292; the CNV2: F(1,13) = 1.586, p > 0.230), temporal length (the CNV1: F (3,39) = 2.705, p > 0.071; except the CNV2: F(3,39) = 4.577, p < 0.019), and the interaction between magnitude and temporal length (the CNV1: F(3,39) = 1.402, p > 0.264; the CNV2: F(3,39) = 1.070, p > 0.365) are observed in the control condition. Besides the above analyses, we also analyzed the mean amplitudes of CNV1 and CNV2 by using the same temporal window (400–500 ms) in the four types of durations. The results are similar to the former analyses. The statistic results of the mean amplitudes of the CNV1 and the CNV2 show that the magnitude has main effect in digit condition (the CNV1: F(1,13) = 4.843, p < 0.047; the CNV2: F (1,13) = 5.290, p < 0.039). It is observed that the two CNV components elicited by small digits are both enhanced than those by large digits. No main effect of temporal length is found (the CNV1: F(3,39) = 1.398, p > 0.264; the CNV2: F(3,39) = 0.585, p>0.584), and no magnitude×temporal length interaction is found (the CNV1: F(3,39) =0.050, p>0.979; the
CNV2: F(3,39)=1.508, p>0.233). No main effect of magnitude (the CNV1: F(1,13)=0.029, p>0.867; the CNV2: F(1,13) =1.424, p>0.254), temporal length (the CNV1: F(3,39)=1.459, p>0.246; except the CNV2: F(3,39) = 3.887, p < 0.020), and the interaction between magnitude and temporal length (the CNV1: F(3,39) = 1.637, p>0.212; the CNV2: F(3,39)=2.555, p>0.082) are observed in the control condition.
2.2.2.
Peak amplitudes of P1s and N1s
P1 and N1 components are closely related to the early stage of perceptual processing (Correa et al., 2005; Correa et al., 2006a; Doherty et al., 2005; Griffin et al., 2002), so we are also
Fig. 3 – (A) Small digits elicited enhanced CNV1 (take the electrode of FCZ as an example) than large digits did with four types of duration (p < 0.012), which is not present for the control condition (p > 0.292). Similar results are also present for CNV2 (for digits, p < 0.038; for control figures, p > 0.230). (B) The scalp distribution map at the initial stage of CNV1 (400–500 ms) for the digit condition.
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –1 42
concerned about the two components. The onset of each visual stimulus elicited P1 and N1 components, peak amplitudes of which were measured at four occipital electrodes (PO5, PO6, PO7, PO8), see Fig. 2A. The ERP results were analyzed by three-way repeated measures ANOVA with magnitude (small and large), temporal length (500 ms, 600 ms, 700 ms and 800 ms) and electrodes (4 electrodes for P1 and N1). All of the P1 and N1 results were subject to the Greenhouse–Geisser epsilon correction. We calculate mean amplitudes of P1s over sliding temporal windows of 20 ms duration with a midpoint varying from 100 ms to 140 ms. The maximum amplitude is taken as P1 peak, which is around 110 ms. The statistics results of P1 peak amplitudes showed that there were no main effects of magnitude for both digit conditions (for the first stimulus, F (1,13) = 0.193, p > 0.667; for the second stimulus, F(1,13) = 0.436, p > 0.520) and control figure conditions (for the first stimulus, F(1,13) = 3.147, p > 0.099; for the second stimulus, F(1,13) = 1.591, p > 0.229). The mean amplitudes of N1s were measured over sliding temporal windows of 20 ms duration with a midpoint varying from 140 ms to 200 ms. The maximum amplitude is taken as N1 peak, which is around 170 ms. The statistic results of N1 peak amplitudes evoked by the first stimulus (F(1,13) = 11.898, p < 0.005) and by the second stimulus (F(1,13) = 5.154, p < 0.041) show that the numeric value has main effect. The N1s elicited by small digits are enhanced than that by large digits (Fig. 4A shows that the N1s evoked to the first digit at PO5), whereas the control figures have no magnitude main effect (for the first stimulus, F(1,13) = 0.010, p > 0.922; for the second stimulus, F(1,13) = 2.331, p > 0.150). Scalp distribution map of small-large difference waves show that the difference does not emerge on 100–120 ms that relate to P1, but on 160– 180 ms that relate to N1 for all of the duration length conditions (Fig. 4B).
3.
Discussion
Accurate timing is a ubiquitous aspect of mental process as well as representing magnitude information in space and quantity (Buhusi and Meck, 2005). The present study showed that the accuracy of temporal comparison is influenced by abstract numeric magnitude. That is, if small digits marked shorter duration and large digits marked longer duration, the accuracy of temporal comparison was higher than the reversed condition, but this trend is not observed in control figures with same perceptual variables. These observations are consistent with our previous results (Xuan et al., 2007), and they also support the idea that magnitude information in time and in numbers is not represented independently by the human brain (Walsh, 2003). Moreover, they imply a generalized and abstract component in the magnitude representation that is irrelevant to any specific physical attribute, such as visual stimuli complexity, luminance, and spatial attention shift etc. Therefore, the present study adds new evidence to what is known about the generalized magnitude system. Among the three parts of the generalized magnitude system, many studies have investigated the space–number relationship in depth. The spatial representation of number, espe-
139
Fig. 4 – (A) No magnitude main effect was present in peak amplitudes of P1s for both digit (p > 0.667) and control (p > 0.099) conditions. However, small digits elicited enhanced N1 than large digits did with four types of durations (p < 0.005), which is not present for the control condition (p > 0.922). (B) The scalp distribution map of the small-large difference waves at P1 period (100–120 ms) and N1 period (160–180 ms) for the digit condition.
cially the small-left, large-right “mental number line”, is already widely accepted by researchers (Dehaene et al., 1993). Recently the representation of time was also investigated. Similar to the Spatial-Numerical Association of Response Codes (SNARC) effect, it was observed that temporal information is represented through left-to-right spatial coordinates (Ishihara et al., 2008; Vallesi et al., 2008). “Early” events are associated with left space, and left-side responses are faster; whereas “late” events are associated with right space, so right-side responses are faster. The existence of “mental time line” or the spatial–temporal association of response codes (STEARC) effect demonstrates the close time–space relationship. In the previous studies on space–number relationship, researchers indicated that Arabic digits automatically activate a magnitude code. The code is represented in terms of left and right parts of an analogical mental number line. Fischer et al. (2003) found that perceiving an irrelevant digit can even influence the allocation of visual spatial attention depending on numeric value. Salillas et al. (2008) further suggested that it is a kind of endogenous and top-down attentional modulation generated by central numbers. It is noted that time also convey magnitude information developing from short (early) to long (late). When duration has to be estimated, elapsing time may be represented progressively from left to right. Since numbers can produce a shift of spatial attention, numbers may also produce a shift of temporal attention from early to late. Under this assumption, it would be easy to understand the present results. Small numbers orient attention to early temporal point, and large numbers direct attention to late temporal point. When participants perceive small numbers, they would generate stronger expectation of event that “time is up”, and tend to judge the duration as “short”; in contrast, the expectation by large
140
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –14 2
numbers is not as strong as that by small numbers, so the duration tends to be judged as “long”. In addition to behavioral performance, we also observed some ERP components associated with the current task. The CNV component, a negative developing slow wave over frontal–parietal central sites, is often observed during the temporal interval between a warning signal and an imperative one. In the present study, we observed amplitudes of both the CNV1 and CNV2 are modulated in a temporal comparison task. One of the possible interpretations was the current CNV results might be related with working memory, since subjects need recall the former duration and compare with the latter in the temporal comparison task. Previous studies suggested that the parietal activity was closely related with magnitude information processing (Pinel et al., 2004), so it was possible that the current CNV results reflect the association between time and number. Besides the above possibilities, CNV is also considered to reflect the level of expectation, preparation, and the sustained attention to the upcoming imperative signal (Mnatsakanian and Tarkka, 2002; Pfeuty et al., 2005; Walter et al., 1964). Jonkman (2006) found that young children showed reduced CNV in a go/nogo ERP study, the results were interpreted as reduced target expectation and motor regulation caused by immature frontoparietal network. CNV is also believed to be related with temporal perception (Macar and Vidal, 2004). In the present study, The amplitudes of both the CNV1 and CNV2 are influenced by numeric magnitude. That is, small digits elicit enhanced CNVs as compared with the way large digits do. The current CNV results accord with the explanations of time attention and the subsequent expectation. Recent studies have demonstrated that the amplitudes of CNVs is sensitive to selective temporal attention: an indicative cue can direct attention resources toward a certain temporal point, and the corresponding CNVs to the cues indicative of shorter intervals tend to be more negative than those indicative of longer intervals. The kinds of change reflect the level of expectation and attention to the upcoming events (Griffin et al., 2002; Miniussi et al., 1999). Los and Heslenfeld (2005) indicated that CNV is used as a covert index of nonspecific preparation. It was found that the CNV is more negative following signals cueing a short foreperiod than a long foreperiod (Vallesi et al., 2007). The current CNV data showed small numbers elicit enhanced CNV, which may be associated with higher level of time attention and stronger expectation to the upcoming event. As compared with larger digits, small digits were regarded as “early” cues, and showed the stronger effect, which might be related with the special phenomenon — “attention reorientation” in the domain of time attention (Miniussi et al., 1999). Whatever the duration was longer or shorter, small digits always induced a stronger expectation to the upcoming event. Even if subjects are cued to an earlier target whereas the ending of duration did not actually appear earlier, subjects can reorient attention to the longer duration. That may be the reason why small digits always induced enhanced CNVs as compared with large digits did. The electrophysiological evidence may help us to understand the underlying mechanisms of numeric influence on the temporal comparison task. It is necessary to note that time attention is only a possible interpretation since we did not
modulate the attention explicitly in the current experiment, which is the limitation of the current study. Besides CNV, N1 is also influenced by numeric value. Similar to the results of CNV, small digits elicit enhanced N1 as compared with the way large digits do. We especially observed different effects on P1 and N1 components in this experiment. Numeric magnitude does not modulate the P1 around 110 ms, but influence the N1s around 170 ms, which make us to further localize the temporal stage of the modulation. These results showed numeric magnitude influenced temporal processing at a relatively early stage of perceptual level that is after the P1 and during the N1 period. One of the possible explanations to the current results is that the N1 data reflects perceptual discrimination between the two numerical categories — small digits and large digits. In previous literature, it was observed that N1 is associated with perceiving distance of numeric magnitude in a number comparison task (Temple and Posner, 1998) or crude estimation of numbers in a parity judgment task (Plodowski et al., 2003). These results suggested that N1 is related with number perceptual discrimination. However, compared with number comparison task in the previous studies, digits are taskirrelevant in the present experiment, so participants need not attend digits actively and need not compare digits intentionally. Moreover, some research also indicated that the perceptual discrimination of number may not be reflected by N1, but P2 or P300 (Schwarz and Heinze, 1998; Turconi et al., 2004). Further experiment is needed to test for whether the present N1 results reflect the perceptual discrimination depending on the specific task requirements of the current study. According to the present data and the above-mentioned analyses, we suggest that temporal attention and subsequent expectation may be possible explanations of the modulation in a temporal comparison task. Small digits are similar to the cues predicting the shorter interval, direct attention to the shorter temporal “point” and elicit enhanced CNV with stronger expectation; in contrast, large digits direct attention to the longer temporal “point”. In other words, digits may play a role of “temporal cue”. Depending upon their numeric magnitudes, digits can flexibly orientate temporal attention to different point that is related with the ending of duration. Then the resulting rational or irrational deployment of temporal attention influence subjective temporal perception. However, it is important to note that time attention is just a possible explanation of the current results. Further studies are needed to validate the role of time attention by direct attentional manipulation.
4.
Experimental procedures
4.1.
Subjects
Fourteen healthy right handed participants (nine males, five females) from the University of Science and Technology of China took part in this experiment. Their age ranged from 22 to 29 years (mean = 24.57 years). All had normal or correctedto-normal vision. All participants gave informed consents and were paid for their participation.
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –1 42
4.2.
Behavioral procedures
Participants sat comfortably in a dimly lit room. The 21q CRT monitor was 140 cm away from them. All procedures were controlled by a personal computer. Response accuracy, reaction time, and EEGs were collected at the same time. Participants were instructed to focus their gaze on the central fixation cross during all the tasks. In each trial, a pair of visual stimuli was presented successively for different duration on the screen center, there was no interval between the presentations of the two durations. Stimuli contained two types, the first type was Arabic digits 1, 2, 3, 4, 6, 7, 8 or 9, with size of 1.6° × 1°. The second type was control figures generated by scrambled images of Arabic digits 1–9, so as to match possible confounding perceptual variables. The first half of digits (1–4) and corresponding controls were labeled as “small”, and the second half of digits (6–9) and corresponding controls as “large” (see Fig. 1A). Duration paired in each trial ranged from 500–600, 600–700 to 700–800 ms (the difference between the shorter and the longer was 100 ms). The shorter or longer durations defined by “small” or “large” stimuli generated two types of experimental conditions: congruent and incongruent. The congruent condition refers to the case in which a “small” stimulus was presented for a shorter time and a “large” stimulus was presented for a longer time. The incongruent condition refers to the “small”/ longer and “larger”/shorter stimulus configuration. According to the configuration of the first stimulus and its presentation of duration in each trial (the second stimulus and its presentation of duration was therefore determined), we can subdivide the congruent and incongruent conditions into four types (digit1/duration1–digit2/duration2): “small”/shorter–“large”/longer, “large”/longer–“small”/shorter, “small”/ longer–“large”/shorter, and “large”/shorter–“small”/longer. Participants were asked to compare the lengths of two durations in each trial, and make their responses on the standard PC keyboard as quickly and accurately as possible. If the former was longer (or shorter), press “F” with left hand; and if the latter was longer (or shorter), press “J” with right hand. The left and right response keys were counterbalanced between participants. Participants were told in advance that visual stimuli per se were irrelevant to temporal tasks. Each subject practiced 32 trials before the formal experiment. There were 4 blocks of 192 trials in the formal experiment.
4.3.
ERP acquisition and analysis
A Neuroscan ESI-128 system was used to record continuous EEG with a cap carry 64 Ag/AgCl electrodes placed at standard locations of the scalp following the extended international 10– 20 system (American Electroencephalographic Society). Two linked referenced electrodes were placed on the bilateral mastoids. Horizontal and vertical bipolar EOGs were monitored with 4 additional electrodes around the eyes. Data was sampled at 1000 Hz and filtered with DC to 100 Hz. Electrodes impedances were kept below 5 kΩ. A DC offset correction was first used offline on the continuous data. Then a FIR low pass filter (zero phase shift, cutoff frequency 20 Hz, roll-off 24 db/octave) was performed. Epochs were started 100 ms before, and ended 1000 ms after
141
the duration onset defined by visual stimuli. Epochs with EOG magnitude beyond ± 50 μv were rejected. Baseline was defined as 100 ms preceding the duration onset. The mean amplitude of baseline was subtracted from the whole epoch to remove baseline offsets in subsequent averages. Finally, epochs were averaged according to event types at all recording channels.
Acknowledgments We thank Reaboka Maraisane for her help with the English. This research was supported by the National Nature Science Foundation of China (30770713, 30870764, and 30800297), the Ministry of Science and Technology of China (2006CB500705) and the Intercollegiate Provincial Nature Science Project of Anhui (KJ2008B78ZC).
REFERENCES
Basso, G., Nichelli, P., Frassinetti, F., di Pellegrino, G., 1996. Time perception in a neglected space. Neuroreport 7, 2111–2114. Buhusi, C.V., Meck, W.H., 2005. What makes us tick? Functional and neural mechanisms of interval timing. Nat. Rev. Neurosci. 6, 755–765. Buonomano, D., Merzenich, M., 1995. Temporal information transformed into a spatial code by a neural network with realistic properties. Science 267, 1028–1030. Burr, D., Morrone, C., 2006. Time perception: space–time in the brain. Curr. Biol. 16, R171–R173. Cohen, J., Hansel, C.E., Sylvester, J.D., 1953. A new phenomenon in time judgment. Nature 172, 901. Correa, A., Lupianez, J., Tudela, P., 2005. Attentional preparation based on temporal expectancy modulates processing at the perceptual level. Psychon. Bull. Rev. 12, 328–334. Correa, A., Lupianez, J., Madrid, E., Tudela, P., 2006a. Temporal attention enhances early visual processing: a review and new evidence from event-related potentials. Brain Res. 1076, 116–128. Correa, A., Sanabria, D., Spence, C., Tudela, P., Lupianez, J., 2006b. Selective temporal attention enhances the temporal resolution of visual perception: evidence from a temporal order judgment task. Brain Res. 1070, 202–205. Coull, J.T., Nobre, A.C., 1998. 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. 18, 7426–7435. Coull, J.T., Vidal, F., Nazarian, B., Macar, F., 2004. Functional anatomy of the attentional modulation of time estimation. Science 303, 1506–1508. Dehaene, S., Bossini, S., Giraux, P., 1993. The mental representation of Parity and Number magnitude. J. Exp. Psychol. General 122, 371–396. Dehaene, S., Dehaene-Lambertz, G., Cohen, L., 1998. Abstract representations of numbers in the animal and human brain. Trends Neurosci. 21, 355–361. DeLong, A.J., 1981. Phenomenological space–time: toward an experiential relativity. Science 213, 681–683. Doherty, J.R., Rao, A., Mesulam, M.M., Nobre, A.C., 2005. Synergistic effect of combined temporal and spatial expectations on visual attention. J. Neurosci. 25, 8259–8266. Doricchi, F., Guariglia, P., Gasparini, M., Tomaiuolo, F., 2005. Dissociation between physical and mental number line bisection in right hemisphere brain damage. Nat. Neurosci. 8, 1663–1665.
142
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –14 2
Dormal, V., Seron, X., Pesenti, M., 2006. Numerosity-duration interference: a Stroop experiment. Acta Psychol. (Amst) 121, 109–124. Fischer, M.H., 2001. Number processing induces spatial performance biases. Neurology 57, 822–826. Fischer, M.H., Castel, A.D., Dodd, M.D., Pratt, J., 2003. Perceiving numbers causes spatial shifts of attention. Nat. Neurosci. 6, 555–556. Griffin, I.C., Miniussi, C., Nobre, A.C., 2002. Multiple mechanisms of selective attention: differential modulation of stimulus processing by attention to space or time. Neuropsychologia 40, 2325–2340. Grondin, S., 1998. Judgments of the duration of visually marked empty time intervals: linking perceived duration and sensitivity. Percept Psychophys. 60, 319–330. Grondin, S., 2001. From physical time to the first and second moments of psychological time. Psychol. Bull. 127, 22–44. Helson, H., 1930. The tau effect—an example of psychological relativity. Science 536–537. Ishihara, M., Keller, P.E., Rossetti, Y., Prinz, W., 2008. Horizontal spatial representations of time: evidence for the STEARC effect. Cortex 44, 454–461. Johnston, A., Arnold, D.H., Nishida, S., 2006. Spatially localized distortions of event time. Curr. Biol. 16, 472–479. Jones, B., Huang, Y., 1982. Space–time dependencies in psychophysical judgment of extent and duration: algebraic models of the tau and kapa effects. Psychol. Bull. 91, 128–142. Jonkman, L.M., 2006. The development of preparation, conflict monitoring and inhibition from early childhood to young adulthood; a Go/Nogo ERP study. Brain Res. 1097, 181–193. Lewis, P.A., Walsh, V., 2005. Time perception: components of the brain's clock. Curr. Biol. 15, R389–R391. Los, S.A., Heslenfeld, D.J., 2005. Intentional and unintentional contributions to nonspecific preparation: electrophysiological evidence. Journal of Experimental Psychology: General 134, 52–72. Macar, F., Grondin, S., Casini, L., 1994. Controlled attention sharing influences time estimation. Mem. Cognit. 22, 673–686. Macar, F., Vidal, F., 2004. Event-related potentials as indices of time processing: a review. J. Psychophysiol. 18, 89–104. Miniussi, C., Wilding, E.L., Coull, J.T., Nobre, A.C., 1999. Orienting attention in time. Modulation of brain potentials. Brain 122 (Pt 8), 1507–1518. Mnatsakanian, E.V., Tarkka, I.M., 2002. Task-specific expectation is revealed in scalp-recorded slow potentials. Brain Topogr. 15, 87–94. Nobre, A.C., 2001. Orienting attention to instants in time. Neuropsychologia 39, 1317–1328. Perbal, S., Couillet, J., Azouvi, P., Pouthas, V., 2003. Relationships between time estimation, memory, attention, and processing speed in patients with severe traumatic brain injury. Neuropsychologia 41, 1599–1610.
Pfeuty, M., Ragot, R., Pouthas, V., 2005. Relationship between CNV and timing of an upcoming event. Neurosci. Lett. 382, 106–111. Pinel, P., Piazza, M., Le Bihan, D., Dehaene, S., 2004. Distributed and overlapping cerebral representations of number, size, and luminance during comparative judgments. Neuron 41, 983–993. Plodowski, A., Swainson, R., Jackson, G.M., Rorden, C., Jackson, S.R., 2003. Mental representation of number in different numerical forms. Curr. Biol. 13, 2045–2050. Posner, M.I., 1980. Orienting of attention. Q. J. Exp. Psychol. 32, 3–25. Pouthas, V., Perbal, S., 2004. Time perception depends on accurate clock mechanisms as well as unimpaired attention and memory processes. Acta Neurobiol. Exp. (Wars) 64, 367–385. Salillas, E., El Yagoubi, R., Semenza, C., 2008. Sensory and cognitive processes of shifts of spatial attention induced by numbers: an ERP study. Cortex 44, 406–413. Schiffman, H.R., Bobko, D.J., 1974. Effects of stimulus complexity on the perception of brief temporal intervals. J. Exp. Psychol. 103, 156–159. Schiffman, H.R., Bobko, D.J., 1977. The role of number and familiarity of stimuli in the perception of brief temporal intervals. Am. J. Psychol. 90, 85–93. Schwarz, W., Heinze, H.J., 1998. On the interaction of numerical and size information in digit comparison: a behavioral and event-related potential study. Neuropsychologia 36, 1167–1179. Sevigny, M.C., Everett, J., Grondin, S., 2003. Depression, attention, and time estimation. Brain Cogn. 53, 351–353. Temple, E., Posner, M.I., 1998. Brain mechanisms of quantity are similar in 5-year-old children and adults. Proc. Natl. Acad. Sci. U. S. A. 95, 7836–7841. Tse, P.U., Intriligator, J., Rivest, J., Cavanagh, P., 2004. Attention and the subjective expansion of time. Percept. Psychophys. 66, 1171–1189. Turconi, E., Jemel, B., Rossion, B., Seron, X., 2004. Electrophysiological evidence for differential processing of numerical quantity and order in humans. Brain Res. Cogn. Brain Res. 21, 22–38. Vallesi, A., Shallice, T., Walsh, V., 2007. Role of the prefrontal cortex in the foreperiod effect: TMS evidence for dual mechanisms in temporal preparation. Cereb. Cortex 17, 466–474. Vallesi, A., Binns, M.A., Shallice, T., 2008. An effect of spatial–temporal association of response codes: understanding the cognitive representations of time. Cognition 107, 501–527. Walsh, V., 2003. A theory of magnitude: common cortical metrics of time, space and quantity. Trends Cogn. Sci. 7, 483–488. Walter, W.G., Cooper, R., Aldridge, V.J., McCallum, W.C., Winter, A.L., 1964. Contingent negative variation: an electric sign of sensorimotor association and expectancy in the human brain. Nature 203, 380–384. Xuan, B., Zhang, D., He, S., Chen, X., 2007. Larger stimuli are judged to last longer. J. Vis. 7, 21–25.
a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m
w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s
Research Report
Numerical magnitude modulates temporal comparison: An ERP study Bin Xuan a,b , Xiang-Chuan Chen a , Sheng He c , Da-Ren Zhang a,⁎ a
Division of Bio-X Interdisciplinary Sciences, Hefei National Laboratory for Physical Sciences at Microscale and Department of Neurobiology and Biophysics, School of Life Science, University of Science and Technology of China, Hefei, Anhui 230027, PR China b Department of Psychology, Anhui Normal University, Wuhu, Anhui 241000, PR China c Department of Psychology, University of Minnesota, 75 East River Road, Minneapolis, MN 55455, USA
A R T I C LE I N FO
AB S T R A C T
Article history:
Time is believed to be a part of the generalized magnitude system just like space and
Accepted 6 March 2009
quantity. Previous research suggests that time perception can be affected by magnitude in
Available online 21 March 2009
some non-temporal dimensions. Here we address two questions. First, could the influence be caused by an abstract magnitude component without perceptual variables? Second, what
Keywords:
are the underlying mechanisms of the influence? Participants compared a pair of durations
Numeric magnitude
defined by two Arabic digits in a hundreds of milliseconds range. They performed more
Temporal comparison
accurately when the shorter durations were defined by lower numeric value digits (small
Selective temporal attention
digits) and the longer durations were defined by higher value digits (large digits) than they
Contingent negative variation
did in the reversed condition. Event-Related Potential (ERP) results showed that the CNVs
N1
corresponding to the first duration (CNV1), to the second duration (CNV2) and the N1 were
Event-Related Potential
all enhanced when durations marked by small digits than that marked by large ones. Combining the electrophysiological data with the behavioral results, we suggest that digits can modulate performance of temporal comparison at the relatively early stage of perceptual processing. One possible explanation of the current results is that selective temporal attention and subsequent expectation may be involved in this modulation. © 2009 Elsevier B.V. All rights reserved.
1.
Introduction
Time is a fundamental dimension of our existence. Processing of temporal information is crucial to many aspects of our daily lives, from our sleep–wake cycle to speech recognition, music perception and skilled motor control, even just waiting an upcoming event (Buhusi and Meck, 2005). We can experience time across a wide range of intervals, but often represent and estimate it subjectively with limited precision, especially in a hundreds of milliseconds range, for its susceptibility to many environmental factors and cognitive processes.
Time is traditionally believed to be a fundamentally different perceptual dimension from space or quantity. However, increasing evidence indicates that perceiving temporal magnitude may not be an independent or specialized neural process, but has close relationships with perceiving other magnitude dimensions in the generalized magnitude system (Walsh, 2003). Take “the Tau effect” as an example, when three points on observer's forearm were stimulated in succession to define two spatial and temporal information, Helson found that the judgments of relative positions were high correlated with the relative intervals. If the temporal interval between
⁎ Corresponding author. Fax: +86 551 3601443. E-mail address: [email protected] (D.R. Zhang). 0006-8993/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2009.03.016
136
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –14 2
stimulating the second and third points was greater than that between the first and second, observers were inclined to report that the tactile spatial distance between the second and third points was greater than that between the first and second, though in fact it might be equal or less (Helson, 1930). Similarly, the judgments of temporal intervals between two stimuli were also reported to depend on their spatial relations (Cohen et al., 1953). It was even observed that the experience of time is compressed together with space in scale-model environments (DeLong, 1981), which reflects the psychologically interdependence between magnitude information in spatial and temporal dimensions. Electrophysiological, neuropsychological and further behavioral studies all implicate the possible spatialrelated timing mechanisms (Basso et al., 1996; Buonomano and Merzenich, 1995; Grondin, 1998; Johnston et al., 2006). On the other hand, the relationships between perceiving time and magnitude in other dimensions were also observed. Dormal et al. found that judging the duration of dot sequences is influenced by the irrelevant number of dots (Dormal et al., 2006). Our previous study also suggest the error rates of temporal judgment could be significantly affected by the magnitudes in non-temporal dimensions, including number of dots, size of open squares, luminance of solid squares, and numeric value of digits (Xuan et al., 2007). In addition, quite a few research also demonstrated that number, another part of the proposed generalized magnitude system, and space are also close related. Dehaene et al. made the pioneer studies of this domain. They put forward “number sense” — a domainspecific, biologically-determined ability (Dehaene et al., 1998), but they also indicated that numbers can be represented and manipulated on a “mental number line” (Dehaene et al., 1993). This analogical spatial representation robustly influences the performance of many spatial cognitive tasks, such as magnitude comparison (Pinel et al., 2004), target detection (Fischer et al., 2003), line and number bisection (Doricchi et al., 2005; Fischer, 2001) etc. Based on the mentioned evidence, Walsh proposed a common magnitude concept. The magnitude concept is rooted in people's need for information about the nature of the external world, and it is often represented as “how many, how much, how long, how far and how fast” with shared metric, processing resources and behavioral goals (Walsh, 2003). However, there still remain disagreements on the relationship between time and magnitude in other dimensions until recently (Lewis and Walsh, 2005). First, some research indicates that short time on the scale of tens to hundreds of milliseconds may not be encoded explicitly as a metric (Burr and Morrone, 2006). Second, it is still a question whether there is an abstract and generalized magnitude component influence time perception. Specifically, the magnitude component is irrelevant to any perceptual variable, such as complexity, luminance, velocity of imputed motion or spatial attention shift (Cohen et al., 1953; Jones and Huang, 1982; Schiffman and Bobko, 1974; Schiffman and Bobko, 1977). To avoid these possible interferential factors, in the present study, Arabic digits were selected to mark durations in a temporal comparison task, and the control stimuli were also used to match any confounding physical parameter. Our first concern focused on whether the subjective time perception is affected by
abstract symbols bearing magnitude meanings such as Arabic digits, and whether the influence is independent of perceptual variables. If digits did modulate the subjective time perception, the underlying mechanisms of the modulation naturally became our next concern. Lots of studies have indicated that time perception is intimately related to the level of attention (Grondin, 2001; Macar et al., 1994; Perbal et al., 2003; Pouthas and Perbal, 2004; Sevigny et al., 2003). Both the accuracy (Coull et al., 2004) and the length (Tse et al., 2004) of subjective time perception can be influenced by attention resource deployment. What is more, Nobre, Coull and their colleagues simultaneously explored two fundamental cognitive functions: selective attention and processing of time. They gave a great impetus to the study of orienting attention to time. Similar to attention in spatial domain, a series of delicate experiments demonstrate that attention in the temporal domain can be directed to certain time point of an upcoming event intentionally, which is named as selective temporal attention (Coull and Nobre, 1998). With a modified Posner's paradigm (Posner, 1980), it was found that subjects can make use of valid temporal cues to correctly predict specific time points in high proportion, effectively improve the performance of target detection and accuracy of predictable temporal information (Correa et al., 2006b; Nobre, 2001). In the mentioned studies, cues were originally meaningless. The meanings of cues were set up through the proportion of correct predictions. However, digits themselves can convey meanings of magnitude. Fischer et al. investigated the use of digits as cues for the detection of lateralized spatial targets. They found that digits can direct spatial attention to the left or right of visual space (Fischer et al., 2003). Similarly, we suspected that digits might play a role of “temporal cue” in the present experiment. That is, digits may direct subjects to deploy attention to a certain time point according to their numerical magnitude. Small digits orient attention to early points in time, and large digits orient attention to late points in time. Based on this hypothesis, if small digits marked shorter duration, large digits marked longer duration, then subjects would establish valid time attention to expect the target — the ending of duration. That is, subjects would expect target arrived earlier with small digits and later with large digits, and it was consistent with this kind of experimental condition; In contrast, if large digits marked shorter duration and small digits marked longer duration, time attention would be invalid. That is, subjects generate a reversed expectation induced by digits as compared with this kind of experimental condition. The valid or invalid time attention would influence the accuracy of temporal comparison. To further validate the hypothesis, we use Event-Related Potential (ERP) to monitor the brain activity. ERP investigation is appropriate to track on-line the dynamic modulation processing of time perception for the high temporal resolution (Macar and Vidal, 2004). If temporal attention did mediate the influence, we can expect that the related ERP components, such as CNV were modulated by digits, since some previous studies indicated that CNV component was related with time attention and the expectation of the upcoming event (Griffin et al., 2002; Miniussi et al., 1999). The combined behavioral performance and the related
137
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –1 42
Fig. 1 – (A) Arabic digits and the control figures (scrambled images of digits) are used to define durations in the experiment. “Small” and “large” were assigned based on digit value. (B) The response accuracy of temporal judgment in the digit (left) and control figure (right) conditions. Subjects tended to judge large digits to last longer (*p < 0.006), which did not happen in control figures. Error bars represent one standard error.
brain potentials would cast light on the influence by digits in a temporal comparison task.
2.
Results
We used two types of stimuli to define duration: Arabic digits and control figures (scrambled digits). See Fig. 1A. There were two durations (one shorter and one longer) defined by two stimuli (one “large” and one “small”) in each trail. Participants were required to compare the lengths of two durations in each trial.
2.1.
Behavioral results
The behavioral performances of 14 subjects, including response accuracy and reaction time were analyzed by two-way repeated measures ANOVA with magnitude (small and large) and duration length (shorter and longer) according to the first stimuli configurations of each trial (see Table 1). There is no duration length main effect (digits: F(1,13)= 1.807, p > 0.202; control figures: F(1,13)= 1.480, p > 0.245) and no magnitude main effect (digits: F
(1,13) = 2.493, p > 0.138; control figures: F(1,13) = 0.630, p > 0.441) in response accuracy. However, there is a significant interaction between duration length and magnitude for digit condition (F (1,13) = 10.843, p < 0.006). That is, the accuracy of temporal comparison in the condition of shorter durations defined by small digits and longer durations defined by large digits is higher than in the reversed condition (see Fig. 1B). We divided the first digit of each trial into eight conditions (1, 2, 3, 4, 6, 7, 8 and 9) in the further detailed analysis. The accuracy of temporal comparison showed a significant magnitude main effect. When the first duration was shorter, accuracy tended to decrease with the increase of numeric value, F(7,91)= 2.884, p < 0.010; when the first duration was longer, accuracy tended to increase with the increase of numeric value, F(7,91) = 5.185, p < 0.0001. However, no similar interaction was found for control condition (F(1,13) = 0.091, p > 0.767). The statistic results of reaction time showed no duration length main effect (digits: F(1,13) = 4.335, p > 0.057; except for control condition: F(1,13) = 20.695, p < 0.001), no magnitude main effect (digits: F(1,13) = 0.567, p > 0.464; control figures: F(1,13)= 0.845, p > 0.374), and no interaction between the two factors (digit: F(1,13)= 1.251, p > 0.283; control figures: F(1,13)= 0.739, p > 0.405).
2.2.
ERP results
2.2.1.
Mean amplitudes of CNVs
Corresponding to the two durations in each trial, the two CNV components were elicited on 400 ms or so after each duration onset. The representative CNV1 (corresponding to the first duration) and CNV2 (corresponding to the second duration) are shown in Fig. 2A (for example, the CNVs elicited by 700–800 duration are shown). According to the previous studies (Correa et al., 2006a; Macar and Vidal, 2004) and our results, electrode grouping was performed in order to define a region of interest (ROI). The ROI includes 9 electrodes of frontal–parietal central site: F1, FZ, F2, FC1, FCZ, FC2, C1, CZ, and C2. The mean amplitudes of CNVs were measured over the four types of temporal window in accordance to the four types of durations. The temporal window ranged from 400 ms to the end of the duration (400–500 ms for duration of 500 ms; 400–600 ms for 600 ms; 400–700 ms for 700 ms; 400–800 ms for 800 ms, see Fig. 2B). Identical analysis was performed on both CNV1 and CNV2.
Table 1 – Response accuracy and reaction time. Digits Accuracy (%) Small Short Long dur (F,p) mag (F,p) dur × mag (F,p)
Large
74.26 62.05 (16.07) (19.75) 66.82 84.48 (17.80) (10.20) 1.807, 0.202 2.493, 0.138 10.843, 0.006
Control figures Reaction time (ms) Small
Large
537.03 559.31 (172.04) (165.66) 608.92 602.55 (151.58) (174.48) 4.335, 0.058 0.567, 0.465 1.251, 0.284
Accuracy (%) Small
Large
67.48 66.74 (15.06) (18.73) 71.35 73.66 (14.25) (14.13) 1.480, 0.245 0.630, 0.441 0.091, 0.768
Reaction time (ms) Small
Large
526.67 527.91 (177.65) (178.19) 622.58 602.62 (158.71) (166.86) 20.695, 0.001 0.845, 0.375 0.739, 0.406
Data in parentheses denote one standard error for all of the four conditions (represented with the first stimuli configuration of each trial: duration1/digit1). Here “dur” means duration length; “mag” means numeric magnitude.
138
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –14 2
Fig. 2 – (A) The red line indicates the representative CNVs at FCZ (including the CNV1 and the CNV2). CNV results were analyzed from the 9 electrodes of the red square. The blue line indicates the representative P1 and N1 components at PO5. The statistics results of P1s and N1s came from the 4 electrodes in the two blue diamonds. (B) The mean amplitudes of CNVs were measured over the temporal window according to the four types of duration: 400–500 ms was used for duration of 500 ms, 400–600 ms for 600 ms, 400–700 ms for 700 ms, and 400–800 ms for 800 ms.
The CNV results were analyzed by three-way repeated measures ANOVA with magnitude (small and large), temporal length (500, 600, 700 and 800 ms) and electrodes (9 electrodes). All of the results were subject to the Greenhouse–Geisser epsilon correction. The statistic results of the mean amplitudes of the CNV1 and the CNV2 show that the magnitude has main effect in digit condition (the CNV1: F(1,13) = 8.723, p < 0.012; the CNV2: F(1,13) = 5.332, p < 0.038). It is observed that the two CNV components elicited by small digits are both enhanced than those by large digits (Fig. 3A). Scalp distribution map shows that the difference in the frontal–parietal area already emerges at the initial stage of CNVs (from 400 to 500 ms) for all of the duration length conditions (Fig. 3B). No main effect of temporal length is found (the CNV1: F(3,39)=1.290, p > 0.289; the CNV2: F(3,39) = 0.304, p > 0.750), and no magnitude × temporal length interaction is found (the CNV1: F(3,39) = 0.127, p > 0.923; the CNV2: F(3,39) = 0.383, p > 0.728). No main effect of magnitude (the CNV1: F(1,13) = 1.204, p > 0.292; the CNV2: F(1,13) = 1.586, p > 0.230), temporal length (the CNV1: F (3,39) = 2.705, p > 0.071; except the CNV2: F(3,39) = 4.577, p < 0.019), and the interaction between magnitude and temporal length (the CNV1: F(3,39) = 1.402, p > 0.264; the CNV2: F(3,39) = 1.070, p > 0.365) are observed in the control condition. Besides the above analyses, we also analyzed the mean amplitudes of CNV1 and CNV2 by using the same temporal window (400–500 ms) in the four types of durations. The results are similar to the former analyses. The statistic results of the mean amplitudes of the CNV1 and the CNV2 show that the magnitude has main effect in digit condition (the CNV1: F(1,13) = 4.843, p < 0.047; the CNV2: F (1,13) = 5.290, p < 0.039). It is observed that the two CNV components elicited by small digits are both enhanced than those by large digits. No main effect of temporal length is found (the CNV1: F(3,39) = 1.398, p > 0.264; the CNV2: F(3,39) = 0.585, p>0.584), and no magnitude×temporal length interaction is found (the CNV1: F(3,39) =0.050, p>0.979; the
CNV2: F(3,39)=1.508, p>0.233). No main effect of magnitude (the CNV1: F(1,13)=0.029, p>0.867; the CNV2: F(1,13) =1.424, p>0.254), temporal length (the CNV1: F(3,39)=1.459, p>0.246; except the CNV2: F(3,39) = 3.887, p < 0.020), and the interaction between magnitude and temporal length (the CNV1: F(3,39) = 1.637, p>0.212; the CNV2: F(3,39)=2.555, p>0.082) are observed in the control condition.
2.2.2.
Peak amplitudes of P1s and N1s
P1 and N1 components are closely related to the early stage of perceptual processing (Correa et al., 2005; Correa et al., 2006a; Doherty et al., 2005; Griffin et al., 2002), so we are also
Fig. 3 – (A) Small digits elicited enhanced CNV1 (take the electrode of FCZ as an example) than large digits did with four types of duration (p < 0.012), which is not present for the control condition (p > 0.292). Similar results are also present for CNV2 (for digits, p < 0.038; for control figures, p > 0.230). (B) The scalp distribution map at the initial stage of CNV1 (400–500 ms) for the digit condition.
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –1 42
concerned about the two components. The onset of each visual stimulus elicited P1 and N1 components, peak amplitudes of which were measured at four occipital electrodes (PO5, PO6, PO7, PO8), see Fig. 2A. The ERP results were analyzed by three-way repeated measures ANOVA with magnitude (small and large), temporal length (500 ms, 600 ms, 700 ms and 800 ms) and electrodes (4 electrodes for P1 and N1). All of the P1 and N1 results were subject to the Greenhouse–Geisser epsilon correction. We calculate mean amplitudes of P1s over sliding temporal windows of 20 ms duration with a midpoint varying from 100 ms to 140 ms. The maximum amplitude is taken as P1 peak, which is around 110 ms. The statistics results of P1 peak amplitudes showed that there were no main effects of magnitude for both digit conditions (for the first stimulus, F (1,13) = 0.193, p > 0.667; for the second stimulus, F(1,13) = 0.436, p > 0.520) and control figure conditions (for the first stimulus, F(1,13) = 3.147, p > 0.099; for the second stimulus, F(1,13) = 1.591, p > 0.229). The mean amplitudes of N1s were measured over sliding temporal windows of 20 ms duration with a midpoint varying from 140 ms to 200 ms. The maximum amplitude is taken as N1 peak, which is around 170 ms. The statistic results of N1 peak amplitudes evoked by the first stimulus (F(1,13) = 11.898, p < 0.005) and by the second stimulus (F(1,13) = 5.154, p < 0.041) show that the numeric value has main effect. The N1s elicited by small digits are enhanced than that by large digits (Fig. 4A shows that the N1s evoked to the first digit at PO5), whereas the control figures have no magnitude main effect (for the first stimulus, F(1,13) = 0.010, p > 0.922; for the second stimulus, F(1,13) = 2.331, p > 0.150). Scalp distribution map of small-large difference waves show that the difference does not emerge on 100–120 ms that relate to P1, but on 160– 180 ms that relate to N1 for all of the duration length conditions (Fig. 4B).
3.
Discussion
Accurate timing is a ubiquitous aspect of mental process as well as representing magnitude information in space and quantity (Buhusi and Meck, 2005). The present study showed that the accuracy of temporal comparison is influenced by abstract numeric magnitude. That is, if small digits marked shorter duration and large digits marked longer duration, the accuracy of temporal comparison was higher than the reversed condition, but this trend is not observed in control figures with same perceptual variables. These observations are consistent with our previous results (Xuan et al., 2007), and they also support the idea that magnitude information in time and in numbers is not represented independently by the human brain (Walsh, 2003). Moreover, they imply a generalized and abstract component in the magnitude representation that is irrelevant to any specific physical attribute, such as visual stimuli complexity, luminance, and spatial attention shift etc. Therefore, the present study adds new evidence to what is known about the generalized magnitude system. Among the three parts of the generalized magnitude system, many studies have investigated the space–number relationship in depth. The spatial representation of number, espe-
139
Fig. 4 – (A) No magnitude main effect was present in peak amplitudes of P1s for both digit (p > 0.667) and control (p > 0.099) conditions. However, small digits elicited enhanced N1 than large digits did with four types of durations (p < 0.005), which is not present for the control condition (p > 0.922). (B) The scalp distribution map of the small-large difference waves at P1 period (100–120 ms) and N1 period (160–180 ms) for the digit condition.
cially the small-left, large-right “mental number line”, is already widely accepted by researchers (Dehaene et al., 1993). Recently the representation of time was also investigated. Similar to the Spatial-Numerical Association of Response Codes (SNARC) effect, it was observed that temporal information is represented through left-to-right spatial coordinates (Ishihara et al., 2008; Vallesi et al., 2008). “Early” events are associated with left space, and left-side responses are faster; whereas “late” events are associated with right space, so right-side responses are faster. The existence of “mental time line” or the spatial–temporal association of response codes (STEARC) effect demonstrates the close time–space relationship. In the previous studies on space–number relationship, researchers indicated that Arabic digits automatically activate a magnitude code. The code is represented in terms of left and right parts of an analogical mental number line. Fischer et al. (2003) found that perceiving an irrelevant digit can even influence the allocation of visual spatial attention depending on numeric value. Salillas et al. (2008) further suggested that it is a kind of endogenous and top-down attentional modulation generated by central numbers. It is noted that time also convey magnitude information developing from short (early) to long (late). When duration has to be estimated, elapsing time may be represented progressively from left to right. Since numbers can produce a shift of spatial attention, numbers may also produce a shift of temporal attention from early to late. Under this assumption, it would be easy to understand the present results. Small numbers orient attention to early temporal point, and large numbers direct attention to late temporal point. When participants perceive small numbers, they would generate stronger expectation of event that “time is up”, and tend to judge the duration as “short”; in contrast, the expectation by large
140
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –14 2
numbers is not as strong as that by small numbers, so the duration tends to be judged as “long”. In addition to behavioral performance, we also observed some ERP components associated with the current task. The CNV component, a negative developing slow wave over frontal–parietal central sites, is often observed during the temporal interval between a warning signal and an imperative one. In the present study, we observed amplitudes of both the CNV1 and CNV2 are modulated in a temporal comparison task. One of the possible interpretations was the current CNV results might be related with working memory, since subjects need recall the former duration and compare with the latter in the temporal comparison task. Previous studies suggested that the parietal activity was closely related with magnitude information processing (Pinel et al., 2004), so it was possible that the current CNV results reflect the association between time and number. Besides the above possibilities, CNV is also considered to reflect the level of expectation, preparation, and the sustained attention to the upcoming imperative signal (Mnatsakanian and Tarkka, 2002; Pfeuty et al., 2005; Walter et al., 1964). Jonkman (2006) found that young children showed reduced CNV in a go/nogo ERP study, the results were interpreted as reduced target expectation and motor regulation caused by immature frontoparietal network. CNV is also believed to be related with temporal perception (Macar and Vidal, 2004). In the present study, The amplitudes of both the CNV1 and CNV2 are influenced by numeric magnitude. That is, small digits elicit enhanced CNVs as compared with the way large digits do. The current CNV results accord with the explanations of time attention and the subsequent expectation. Recent studies have demonstrated that the amplitudes of CNVs is sensitive to selective temporal attention: an indicative cue can direct attention resources toward a certain temporal point, and the corresponding CNVs to the cues indicative of shorter intervals tend to be more negative than those indicative of longer intervals. The kinds of change reflect the level of expectation and attention to the upcoming events (Griffin et al., 2002; Miniussi et al., 1999). Los and Heslenfeld (2005) indicated that CNV is used as a covert index of nonspecific preparation. It was found that the CNV is more negative following signals cueing a short foreperiod than a long foreperiod (Vallesi et al., 2007). The current CNV data showed small numbers elicit enhanced CNV, which may be associated with higher level of time attention and stronger expectation to the upcoming event. As compared with larger digits, small digits were regarded as “early” cues, and showed the stronger effect, which might be related with the special phenomenon — “attention reorientation” in the domain of time attention (Miniussi et al., 1999). Whatever the duration was longer or shorter, small digits always induced a stronger expectation to the upcoming event. Even if subjects are cued to an earlier target whereas the ending of duration did not actually appear earlier, subjects can reorient attention to the longer duration. That may be the reason why small digits always induced enhanced CNVs as compared with large digits did. The electrophysiological evidence may help us to understand the underlying mechanisms of numeric influence on the temporal comparison task. It is necessary to note that time attention is only a possible interpretation since we did not
modulate the attention explicitly in the current experiment, which is the limitation of the current study. Besides CNV, N1 is also influenced by numeric value. Similar to the results of CNV, small digits elicit enhanced N1 as compared with the way large digits do. We especially observed different effects on P1 and N1 components in this experiment. Numeric magnitude does not modulate the P1 around 110 ms, but influence the N1s around 170 ms, which make us to further localize the temporal stage of the modulation. These results showed numeric magnitude influenced temporal processing at a relatively early stage of perceptual level that is after the P1 and during the N1 period. One of the possible explanations to the current results is that the N1 data reflects perceptual discrimination between the two numerical categories — small digits and large digits. In previous literature, it was observed that N1 is associated with perceiving distance of numeric magnitude in a number comparison task (Temple and Posner, 1998) or crude estimation of numbers in a parity judgment task (Plodowski et al., 2003). These results suggested that N1 is related with number perceptual discrimination. However, compared with number comparison task in the previous studies, digits are taskirrelevant in the present experiment, so participants need not attend digits actively and need not compare digits intentionally. Moreover, some research also indicated that the perceptual discrimination of number may not be reflected by N1, but P2 or P300 (Schwarz and Heinze, 1998; Turconi et al., 2004). Further experiment is needed to test for whether the present N1 results reflect the perceptual discrimination depending on the specific task requirements of the current study. According to the present data and the above-mentioned analyses, we suggest that temporal attention and subsequent expectation may be possible explanations of the modulation in a temporal comparison task. Small digits are similar to the cues predicting the shorter interval, direct attention to the shorter temporal “point” and elicit enhanced CNV with stronger expectation; in contrast, large digits direct attention to the longer temporal “point”. In other words, digits may play a role of “temporal cue”. Depending upon their numeric magnitudes, digits can flexibly orientate temporal attention to different point that is related with the ending of duration. Then the resulting rational or irrational deployment of temporal attention influence subjective temporal perception. However, it is important to note that time attention is just a possible explanation of the current results. Further studies are needed to validate the role of time attention by direct attentional manipulation.
4.
Experimental procedures
4.1.
Subjects
Fourteen healthy right handed participants (nine males, five females) from the University of Science and Technology of China took part in this experiment. Their age ranged from 22 to 29 years (mean = 24.57 years). All had normal or correctedto-normal vision. All participants gave informed consents and were paid for their participation.
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –1 42
4.2.
Behavioral procedures
Participants sat comfortably in a dimly lit room. The 21q CRT monitor was 140 cm away from them. All procedures were controlled by a personal computer. Response accuracy, reaction time, and EEGs were collected at the same time. Participants were instructed to focus their gaze on the central fixation cross during all the tasks. In each trial, a pair of visual stimuli was presented successively for different duration on the screen center, there was no interval between the presentations of the two durations. Stimuli contained two types, the first type was Arabic digits 1, 2, 3, 4, 6, 7, 8 or 9, with size of 1.6° × 1°. The second type was control figures generated by scrambled images of Arabic digits 1–9, so as to match possible confounding perceptual variables. The first half of digits (1–4) and corresponding controls were labeled as “small”, and the second half of digits (6–9) and corresponding controls as “large” (see Fig. 1A). Duration paired in each trial ranged from 500–600, 600–700 to 700–800 ms (the difference between the shorter and the longer was 100 ms). The shorter or longer durations defined by “small” or “large” stimuli generated two types of experimental conditions: congruent and incongruent. The congruent condition refers to the case in which a “small” stimulus was presented for a shorter time and a “large” stimulus was presented for a longer time. The incongruent condition refers to the “small”/ longer and “larger”/shorter stimulus configuration. According to the configuration of the first stimulus and its presentation of duration in each trial (the second stimulus and its presentation of duration was therefore determined), we can subdivide the congruent and incongruent conditions into four types (digit1/duration1–digit2/duration2): “small”/shorter–“large”/longer, “large”/longer–“small”/shorter, “small”/ longer–“large”/shorter, and “large”/shorter–“small”/longer. Participants were asked to compare the lengths of two durations in each trial, and make their responses on the standard PC keyboard as quickly and accurately as possible. If the former was longer (or shorter), press “F” with left hand; and if the latter was longer (or shorter), press “J” with right hand. The left and right response keys were counterbalanced between participants. Participants were told in advance that visual stimuli per se were irrelevant to temporal tasks. Each subject practiced 32 trials before the formal experiment. There were 4 blocks of 192 trials in the formal experiment.
4.3.
ERP acquisition and analysis
A Neuroscan ESI-128 system was used to record continuous EEG with a cap carry 64 Ag/AgCl electrodes placed at standard locations of the scalp following the extended international 10– 20 system (American Electroencephalographic Society). Two linked referenced electrodes were placed on the bilateral mastoids. Horizontal and vertical bipolar EOGs were monitored with 4 additional electrodes around the eyes. Data was sampled at 1000 Hz and filtered with DC to 100 Hz. Electrodes impedances were kept below 5 kΩ. A DC offset correction was first used offline on the continuous data. Then a FIR low pass filter (zero phase shift, cutoff frequency 20 Hz, roll-off 24 db/octave) was performed. Epochs were started 100 ms before, and ended 1000 ms after
141
the duration onset defined by visual stimuli. Epochs with EOG magnitude beyond ± 50 μv were rejected. Baseline was defined as 100 ms preceding the duration onset. The mean amplitude of baseline was subtracted from the whole epoch to remove baseline offsets in subsequent averages. Finally, epochs were averaged according to event types at all recording channels.
Acknowledgments We thank Reaboka Maraisane for her help with the English. This research was supported by the National Nature Science Foundation of China (30770713, 30870764, and 30800297), the Ministry of Science and Technology of China (2006CB500705) and the Intercollegiate Provincial Nature Science Project of Anhui (KJ2008B78ZC).
REFERENCES
Basso, G., Nichelli, P., Frassinetti, F., di Pellegrino, G., 1996. Time perception in a neglected space. Neuroreport 7, 2111–2114. Buhusi, C.V., Meck, W.H., 2005. What makes us tick? Functional and neural mechanisms of interval timing. Nat. Rev. Neurosci. 6, 755–765. Buonomano, D., Merzenich, M., 1995. Temporal information transformed into a spatial code by a neural network with realistic properties. Science 267, 1028–1030. Burr, D., Morrone, C., 2006. Time perception: space–time in the brain. Curr. Biol. 16, R171–R173. Cohen, J., Hansel, C.E., Sylvester, J.D., 1953. A new phenomenon in time judgment. Nature 172, 901. Correa, A., Lupianez, J., Tudela, P., 2005. Attentional preparation based on temporal expectancy modulates processing at the perceptual level. Psychon. Bull. Rev. 12, 328–334. Correa, A., Lupianez, J., Madrid, E., Tudela, P., 2006a. Temporal attention enhances early visual processing: a review and new evidence from event-related potentials. Brain Res. 1076, 116–128. Correa, A., Sanabria, D., Spence, C., Tudela, P., Lupianez, J., 2006b. Selective temporal attention enhances the temporal resolution of visual perception: evidence from a temporal order judgment task. Brain Res. 1070, 202–205. Coull, J.T., Nobre, A.C., 1998. 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. 18, 7426–7435. Coull, J.T., Vidal, F., Nazarian, B., Macar, F., 2004. Functional anatomy of the attentional modulation of time estimation. Science 303, 1506–1508. Dehaene, S., Bossini, S., Giraux, P., 1993. The mental representation of Parity and Number magnitude. J. Exp. Psychol. General 122, 371–396. Dehaene, S., Dehaene-Lambertz, G., Cohen, L., 1998. Abstract representations of numbers in the animal and human brain. Trends Neurosci. 21, 355–361. DeLong, A.J., 1981. Phenomenological space–time: toward an experiential relativity. Science 213, 681–683. Doherty, J.R., Rao, A., Mesulam, M.M., Nobre, A.C., 2005. Synergistic effect of combined temporal and spatial expectations on visual attention. J. Neurosci. 25, 8259–8266. Doricchi, F., Guariglia, P., Gasparini, M., Tomaiuolo, F., 2005. Dissociation between physical and mental number line bisection in right hemisphere brain damage. Nat. Neurosci. 8, 1663–1665.
142
BR A I N R ES E A RC H 1 2 6 9 ( 2 00 9 ) 1 3 5 –14 2
Dormal, V., Seron, X., Pesenti, M., 2006. Numerosity-duration interference: a Stroop experiment. Acta Psychol. (Amst) 121, 109–124. Fischer, M.H., 2001. Number processing induces spatial performance biases. Neurology 57, 822–826. Fischer, M.H., Castel, A.D., Dodd, M.D., Pratt, J., 2003. Perceiving numbers causes spatial shifts of attention. Nat. Neurosci. 6, 555–556. Griffin, I.C., Miniussi, C., Nobre, A.C., 2002. Multiple mechanisms of selective attention: differential modulation of stimulus processing by attention to space or time. Neuropsychologia 40, 2325–2340. Grondin, S., 1998. Judgments of the duration of visually marked empty time intervals: linking perceived duration and sensitivity. Percept Psychophys. 60, 319–330. Grondin, S., 2001. From physical time to the first and second moments of psychological time. Psychol. Bull. 127, 22–44. Helson, H., 1930. The tau effect—an example of psychological relativity. Science 536–537. Ishihara, M., Keller, P.E., Rossetti, Y., Prinz, W., 2008. Horizontal spatial representations of time: evidence for the STEARC effect. Cortex 44, 454–461. Johnston, A., Arnold, D.H., Nishida, S., 2006. Spatially localized distortions of event time. Curr. Biol. 16, 472–479. Jones, B., Huang, Y., 1982. Space–time dependencies in psychophysical judgment of extent and duration: algebraic models of the tau and kapa effects. Psychol. Bull. 91, 128–142. Jonkman, L.M., 2006. The development of preparation, conflict monitoring and inhibition from early childhood to young adulthood; a Go/Nogo ERP study. Brain Res. 1097, 181–193. Lewis, P.A., Walsh, V., 2005. Time perception: components of the brain's clock. Curr. Biol. 15, R389–R391. Los, S.A., Heslenfeld, D.J., 2005. Intentional and unintentional contributions to nonspecific preparation: electrophysiological evidence. Journal of Experimental Psychology: General 134, 52–72. Macar, F., Grondin, S., Casini, L., 1994. Controlled attention sharing influences time estimation. Mem. Cognit. 22, 673–686. Macar, F., Vidal, F., 2004. Event-related potentials as indices of time processing: a review. J. Psychophysiol. 18, 89–104. Miniussi, C., Wilding, E.L., Coull, J.T., Nobre, A.C., 1999. Orienting attention in time. Modulation of brain potentials. Brain 122 (Pt 8), 1507–1518. Mnatsakanian, E.V., Tarkka, I.M., 2002. Task-specific expectation is revealed in scalp-recorded slow potentials. Brain Topogr. 15, 87–94. Nobre, A.C., 2001. Orienting attention to instants in time. Neuropsychologia 39, 1317–1328. Perbal, S., Couillet, J., Azouvi, P., Pouthas, V., 2003. Relationships between time estimation, memory, attention, and processing speed in patients with severe traumatic brain injury. Neuropsychologia 41, 1599–1610.
Pfeuty, M., Ragot, R., Pouthas, V., 2005. Relationship between CNV and timing of an upcoming event. Neurosci. Lett. 382, 106–111. Pinel, P., Piazza, M., Le Bihan, D., Dehaene, S., 2004. Distributed and overlapping cerebral representations of number, size, and luminance during comparative judgments. Neuron 41, 983–993. Plodowski, A., Swainson, R., Jackson, G.M., Rorden, C., Jackson, S.R., 2003. Mental representation of number in different numerical forms. Curr. Biol. 13, 2045–2050. Posner, M.I., 1980. Orienting of attention. Q. J. Exp. Psychol. 32, 3–25. Pouthas, V., Perbal, S., 2004. Time perception depends on accurate clock mechanisms as well as unimpaired attention and memory processes. Acta Neurobiol. Exp. (Wars) 64, 367–385. Salillas, E., El Yagoubi, R., Semenza, C., 2008. Sensory and cognitive processes of shifts of spatial attention induced by numbers: an ERP study. Cortex 44, 406–413. Schiffman, H.R., Bobko, D.J., 1974. Effects of stimulus complexity on the perception of brief temporal intervals. J. Exp. Psychol. 103, 156–159. Schiffman, H.R., Bobko, D.J., 1977. The role of number and familiarity of stimuli in the perception of brief temporal intervals. Am. J. Psychol. 90, 85–93. Schwarz, W., Heinze, H.J., 1998. On the interaction of numerical and size information in digit comparison: a behavioral and event-related potential study. Neuropsychologia 36, 1167–1179. Sevigny, M.C., Everett, J., Grondin, S., 2003. Depression, attention, and time estimation. Brain Cogn. 53, 351–353. Temple, E., Posner, M.I., 1998. Brain mechanisms of quantity are similar in 5-year-old children and adults. Proc. Natl. Acad. Sci. U. S. A. 95, 7836–7841. Tse, P.U., Intriligator, J., Rivest, J., Cavanagh, P., 2004. Attention and the subjective expansion of time. Percept. Psychophys. 66, 1171–1189. Turconi, E., Jemel, B., Rossion, B., Seron, X., 2004. Electrophysiological evidence for differential processing of numerical quantity and order in humans. Brain Res. Cogn. Brain Res. 21, 22–38. Vallesi, A., Shallice, T., Walsh, V., 2007. Role of the prefrontal cortex in the foreperiod effect: TMS evidence for dual mechanisms in temporal preparation. Cereb. Cortex 17, 466–474. Vallesi, A., Binns, M.A., Shallice, T., 2008. An effect of spatial–temporal association of response codes: understanding the cognitive representations of time. Cognition 107, 501–527. Walsh, V., 2003. A theory of magnitude: common cortical metrics of time, space and quantity. Trends Cogn. Sci. 7, 483–488. Walter, W.G., Cooper, R., Aldridge, V.J., McCallum, W.C., Winter, A.L., 1964. Contingent negative variation: an electric sign of sensorimotor association and expectancy in the human brain. Nature 203, 380–384. Xuan, B., Zhang, D., He, S., Chen, X., 2007. Larger stimuli are judged to last longer. J. Vis. 7, 21–25.