Conceived and designed the experiments: ML AS TBP. Performed the experiments: ML AS. Analyzed the data: ML AS. Contributed reagents/materials/analysis tools: ML AS. Wrote the paper: AS ML TBP.
The authors have declared that no competing interests exist.
Emotions change our perception of time. In the past, this has been attributed primarily to emotions speeding up an “internal clock” thereby increasing subjective time estimates. Here we probed this account using an S1/S2 temporal discrimination paradigm. Participants were presented with a stimulus (S1) followed by a brief delay and then a second stimulus (S2) and indicated whether S2 was shorter or longer in duration than S1. We manipulated participants' emotions by presenting a task-irrelevant picture following S1 and preceding S2. Participants were more likely to judge S2 as shorter than S1 when the intervening picture was emotional as compared to neutral. This effect held independent of S1 and S2 modality (Visual: Exps. 1, 2, & 3; Auditory: Exp. 4) and intervening picture valence (Negative: Exps. 1, 2 & 4; Positive: Exp. 3). Moreover, it was replicated in a temporal reproduction paradigm (Exp. 5) where a timing stimulus was preceded by an emotional or neutral picture and participants were asked to reproduce the duration of the timing stimulus. Taken together, these findings indicate that emotional experiences may decrease temporal estimates and thus raise questions about the suitability of internal clock speed explanations of emotion effects on timing. Moreover, they highlight attentional mechanisms as a viable alternative.
In an effort to distinguish man from beast, ancient philosophers dissociated thought from passion and attributed thought to the workings of the brain and passion to the workings of other bodily organs
Amidst the growth of studies on the interaction between emotion and cognition, insights into the role of emotions on interval timing have remained limited. Interval timing typically refers to our capacity to monitor and store durations in the hundreds of milliseconds to minutes range and to adjust our behavior based on these durations. For example, we may have a stored temporal representation of how long it takes for someone to answer a knock at their door. Thus, if a person fails to open the door within that time period, we will not continue knocking indefinitely, but will instead move on once the evoked temporal representation has elapsed. Sensitivity to time is a fundamental aspect of many higher order cognitive processes including decision making, memory and language (for a review see
An indication of a role for emotions in time perception comes from both linguistics and timing experiments. The saying “time drags” is typically used in the context of negative or boring events, whereas the saying “time flies” is used in the context of positive or exciting events
While pacemaker rate
Evidence for the later proposition comes from research that highlights the capacity of emotional stimuli to capture and hold attention. For example, visual search studies indicate that participants are faster at detecting an emotional target among neutral distracters and slower at detecting a neutral target among emotional distracters relative to an all neutral baseline
While there is ample evidence of participants overestimating the duration of emotional relative to neutral events
Fourteen female college students participated in the present study in fulfillment of an introductory psychology course requirement. Their mean age was 19.6 years (SD = 0.65). They were all right-handed
A filled circle 1.7 cm in diameter was used as the timing stimulus (i.e., S1 and S2). Sixty-four task-irrelevant images were selected from the International Affective Picture System (IAPS
Participants were seated in front of a computer screen, instructed on their task and subsequently commenced the experiment. Each trial in the experiment began with the presentation of a fixation cross at screen center for 500 ms followed by a 500 ms blank screen. Then, a circle (S1) appeared for 1200 ms followed by another blank screen of 1400, 1600, 1800, or 2000 ms duration. Next, the participant saw a negative or a neutral picture for 800 ms, followed by a 500 ms blank screen and another circle (S2). S2 was either shorter (i.e., 1040 or 1120 ms) or longer (i.e., 1280 or 1360 ms) than S1. Participants were informed that distracter pictures interleaved every S1/S2 pair and were asked to simply keep looking at the screen. The task instruction was to compare the duration of the two circles (S1 and S2) and to indicate whether S2 was shorter or longer than S1 by pressing one of two buttons on a response box following S2 offset. A new trial started 6300, 6600, 6900, or 7200 ms after the offset of S2 (
There were 64 trials in the experiment. Each S2 duration was presented 8 times in the negative condition and 8 times in the neutral emotion condition. Trial order was randomized and the assignment of response buttons to response alternatives was counterbalanced across participants.
Probability of responding “shorter” was subjected to an ANOVA with Emotion (negative, neutral) and S2 Duration (4 levels) as repeated measures factors. There were significant main effects of S2 Duration (
In contrast to previous studies, we found that participants underestimated stimulus duration during emotional as compared to neutral trials. Specifically, participants in the present experiment were more likely to indicate that S2 duration was shorter than that of S1 when S2 was preceded by a negative as compared to neutral picture. Given the nature of the present paradigm, one may speculate that attention, rather than pacemaker rate, accounts for the observed effects. Emotional modulation of attention may have caused participants to accumulate fewer pacemaker pulses such that perceived S2 duration was reduced. However, before such a conclusion can be drawn, alternative accounts require consideration. Specifically, it is possible that the relatively short interval between the picture and S2 accounts for the observed effects. Perhaps the negative picture engaged perceptual mechanisms that led to a transient distraction causing the participant to miss S2 onset.
Evidence for this possibility comes from research employing the “attentional blink” paradigm. In this paradigm, participants are presented with a stream of distracters and one or two targets. Participants are likely to miss a second target if it closely follows the first target
Fifteen female college students participated in the present study in fulfillment of an introductory psychology course requirement. Their mean age was 18.9 years (SD = 0.64). They were all right-handed
In Experiment 2, a blank computer screen was presented for 2000 ms between the offset of the picture and the onset of S2. Nevertheless, the average trial duration was comparable to Experiment 1 as the intervals between trials were shortened to 5300 and 6200 ms. All other parameters were left unchanged.
Analysis of the probability of responding “shorter” revealed significant main effects of S2 Duration (
The results of Experiment 2 are similar to those of Experiment 1. Thus, we can exclude the possibility that the effects elicited in Experiment 1 reflect an attentional blink following negative pictures. Instead, we argue that the negative pictures triggered emotional processes that distracted individuals from monitoring the S2 duration, leading to a loss of pacemaker pulses and a temporal underestimation relative to the neutral condition. In a third experiment, we sought to replicate this observation and to determine whether it is specific to negative emotions. Past research revealed consistent evidence for attention capture and distraction by negative emotional events
Thirteen female college students participated in the present study in fulfillment of an introductory psychology course requirement. Their mean age was 20.2 years (SD = 1.1). They were all right-handed
Experiment 3 was equivalent to Experiment 1 with the exception that the negative IAPS pictures were replaced by positive IAPS pictures. Examples of positive pictures are erotic couple, fireworks, and roller coaster scenes. Positive pictures had a mean arousal rating of 5.72 (range 5.08–6.99) and a mean valence rating of 7.29 (range 6.56–8.22). The mean arousal rating of the positive pictures used in Experiment 3 was significantly lower (
Analysis of the probability of responding “shorter” revealed significant main effects of S2 Duration (
A direct comparison between positive and negative conditions was conducted by subjecting the data from Experiments 1 and 3 to a 3-way Experiment (1 vs. 3)×Emotion (emotional vs. neutral)×S2 Duration (4 levels) mixed design ANOVA with probability of responding “shorter” as the dependent variable. There was a significant S2 Duration effect (
Finally, we explored the contribution of picture valence and arousal to the observed effects in Experiment 1 and 3. IAPS valence scores were converted to emotion intensity scores by subtracting 5 (the rating scale midpoint) from each value and taking the absolute value of the result. A correlation analysis for these and the IAPS arousal scores revealed that both measures were highly correlated (r = .94, p<.0001). We subsequently tested two ANOVA models to determine whether and in what way each measure predicted the probability of responding “shorter” in Experiments 1 and 3. The first model included Emotional Intensity and Experiment as factors and returned a significant effect of Emotional Intensity (F(1,124) = 5.6, p<.05) with all other effects being non-significant (p>.1). The second model included Arousal and Experiment as factors and returned a significant effect of Arousal (F(1,124) = 5.2, p<.05) with all other effects being non-significant (p>.1).
Participants perceived the duration of a stimulus to be shorter when it was preceded by a positive as compared to a neutral picture. Given that this effect was comparable to that observed for negative pictures, one may conclude that positive and negative stimuli similarly affect S2 duration estimation. S2 is perceived to be shorter when immediately preceded by emotional as compared to neutral stimuli. A discussion of non-specific emotion mechanisms by which this effect may occur is presented in the General Discussion.
Below we report a fourth experiment aimed at addressing the question of whether emotion induced changes in attention to time are limited to within modality processing or persist when emotion and timing information come from different modalities. This question is important as attention is a limited resource
Eighteen female college students participated in the present study in fulfillment of an introductory psychology course requirement. Their mean age was 19.9 years (SD = 0.77). They were all right-handed
Experiment 4 was equivalent to Experiment 1, with the exception that the timing stimulus (i.e., S1/S2) was a 500 Hz pure tone rather than a circle.
Analysis of the probability of responding “shorter” revealed significant main effects of S2 Duration (
In line with Experiments 1–3, participants underestimated the duration of S2 when it was preceded by an emotional as compared to a neutral picture. Given that Experiment 4 used an auditory timing stimulus and manipulated participant emotions through visual stimuli, these results suggest that the effects of emotion on time perception are not limited to within-modality processes, but extend to across modality processes.
Together the current results are opposite to those typically reported in the literature. Emotions here led to temporal underestimation rather than overestimations thus making an attention account a viable competitor of the existing pacemaker account. Nevertheless, one may object that the divergence in results stems from a divergence in methodology, which employed an S1–S2 paradigm in the present case as opposed to the bisection or temporal reproduction paradigms in past research. Therefore, emotion effects on responses may not reflect a shortening of S2 but a lengthening of S1, which was being encoded and maintained in working memory during the emotional challenge. To address this possibility we conducted a final experiment in which each trial comprised the presentation of an emotional or a neutral picture followed by a sample duration, which participants were required to reproduce. If emotions increase the sense of passed and passing time, participants should produce longer durations when primed with an emotional as compared to a neutral image. If, however, emotions distract from the monitoring of time when dissociated from the timing signal, then participants should produce shorter durations when primed with an emotional as compared to a neutral image.
Eighteen female college students participated in the present study in fulfillment of an introductory psychology course requirement. The data of four participants was discarded because they were unable to accurately discriminate the durations presented for reproduction. The mean age of the remaining participants was 22.4 years (SD = 1.86). They were all right-handed
A trial started with the presentation of a fixation cross, which remained on screen for 500 ms. The cross was followed by a picture presentation for 800 ms. After a short empty interval of 500 ms, a circle appeared on the screen for 1100, 1700 or 2300 ms. After a longer empty interval of 5500 ms, a 500 ms cross prompted participants to attend to the onset of another circle. Participants were asked to push a button when they thought the duration of the second circle matched that of the first circle. Inter-trial intervals were equally distributed among 1800, 2100, and 2400 ms.
The visual stimuli used here were comparable to those of Experiment 1. On half the trials, a timing stimulus was preceded by an emotionally negative picture, whereas on the remaining trials it was preceded by a neutral picture. There were a total of 30 emotional and 30 neutral trials with one third each for the 1100, 1700 and 2300 ms duration conditions. Trials were presented in random order.
We computed mean reaction times and standard deviations for each participant across all experimental conditions and removed trials from the analysis in which participants deviated more than two standard deviations from their respective mean. Using the remaining trials, we then computed means for each participant and condition and subjected those to an ANOVA with Emotion (negative, neutral) and Duration (1100, 1700, 2300) as repeated measures factors. This analysis revealed an Emotion main effect (
The subjective shortening of duration when the timing stimulus was preceded by an emotional stimulus in the reproduction paradigm of Experiment 5 confirms that the results of Experiments 1–4 were not merely an artifact of the S1/S2 discrimination paradigm. Moreover, Experiment 5 eliminates the possibility that they were due to emotional events lengthening duration representations that are concurrently maintained in working memory. Rather, emotional events seem to affect temporal processing prospectively. Compared to neutral events, emotional events shorten the perception of subsequently presented timing signals.
The present study set out to investigate the influence of emotions on time perception. Of particular interest was whether pacemaker rate or attention represents a more viable mechanism for emotional effects on timing. Across five experiments, we found that participants perceived a timing stimulus as shorter when it was preceded by an emotional as compared to a neutral distracter. Moreover, this effect was observed for within modality and across modality manipulations, for positively and negatively valenced emotional distracters, for 500 ms and 2000 ms intervals between the distracter and the timing stimulus, and was independent of the type of timing response (i.e., temporal discrimination or reproduction) participants made.
Together these results challenge the idea that emotions influence time perception merely by influencing pacemaker rate. If that were the case, then the present study should have replicated prior work indicating that participants overestimate the duration of emotional as compared to neutral events
First, it is possible that participants directed greater attention to the encoding of emotional as compared to neutral pictures and thus tended to miss the onset of the timing stimulus. In support of this, we found the emotion effect to be statistically independent of timing stimulus duration. However, the durations used here may not have been different enough and/or the statistical power too small to detect duration related changes in the emotion effect. In support of this, we found that emotion induced underestimation was unaffected by an increase in the interval between the emotion stimulus and the timing stimulus in Experiment 2. This increase should have reduced an underestimation effect due to a delayed latency to begin timing because participants had more time to recover their attention. Moreover, in Experiment 5 participants reproduced shorter durations in the emotion condition despite the fact that they were prompted about the impending reproduction of the timing stimulus. Finally, the condition means in the latter experiment suggest that the temporal shortening effect increases from the short (56 ms), to the intermediate (61 ms) to the long duration condition (91 ms). Thus, there are likely other attentional mechanisms contributing to the present results besides a mere onset effect.
One such mechanism could involve the effect of non-specific bodily arousal on attention. Previous research has indicated a relationship between bodily arousal and cognitive performance, which is commonly known as the Yerkes-Dodson law
A final mechanism we would like to propose involves attention capture by appraisal processes and/or emotion regulation attempts
Although our data support an attention account, one may object that a pacemaker rate account cannot be fully ruled out. This is because participants may have used the emotional/neutral picture as a reference for judging or reproducing the target duration. As emotional stimuli were found to be overestimated in previous studies, a longer timing reference may have led to an underestimation of the target. However, the following consideration makes this possibility implausible. Changes in arousal after an emotional challenge develop in time and increases in arousal markers such as heart rate and skin conductance have been noted to peak between 3 to 6 seconds following stimulus onset
Although the pacemaker account has been very popular as an explanation of the relationship between emotion and time, the present work is not the first to suggest a relationship between emotion, attention, and time. For example, Meck and MacDonald
Although prior work cited above raises the possibility that attention capture or aversion modulate the relationship between emotion and time, it does not exclude a pacemaker account. Apart from modulating attention, lesions to the amygdala are known to impair arousal responses to emotional stimuli
The present set of experiments demonstrates that individuals underestimate the duration of a neutral stimulus if that stimulus is presented in an emotional as compared to neutral context. This is true regardless of whether the emotion is positive or negative suggesting that both positive and negative experiences can affect timing in similar ways. The mechanism by which this occurs is unlikely to involve the pacemaker, but instead may arise from a modulation of attention. If true, this would lend a parsimonious explanation of why emotions can both slow down and speed up everyday experiences. A slowing down of perceived time may occur when individuals wish for an experience to end quickly and hence direct resources towards monitoring the passing of time. In contrast, a speeding up of time occurs when the experience captures attention leaving the passing of time unnoticed and leading us to believe that “time flies”.