Visual Duration but Not Numerosity Is Distorted While Running

Abstract

There is increasing evidence that action and perception interact in the processing of magnitudes such as duration and numerosity. Sustained physical exercise (such as running or cycling) increases the apparent duration of visual stimuli presented during the activity. However, the effect of exercise on numerosity perception has not yet been investigated. Here, we asked participants to make either a temporal or a numerical judgment by comparing the duration or numerosity of standard stimuli displayed at rest with those presented while running. The results support previous reports in showing that physical activity significantly expands perceived duration; however, it had no effect on perceived numerosity. Furthermore, the distortions of the perceived durations vanished soon after the running session, making it unlikely that physiological factors such as heart rate underlie the temporal distortion. Taken together, these results suggest a domain-selective influence of the motor system on the perception of time, rather than a general effect on magnitude.

Keywords: atom; magnitude system; numerosity perception; physical activity; time perception.

Figure 1

Experimental procedure. (A) Paradigm used to measure duration and numerosity perception while running. After a short training session, participants were presented with the reference stimulus (encoding phase: 600 ms or 24 dots) 5 times. After a rest period of 3 min, they were presented with a sequence of test stimuli to categorize as same or different, compared to the reference (baseline T1). After the task, the encoding was repeated, followed by the running phase. During the first three minutes of running, no stimuli were presented; then, the same different task (duration or numerosity) was performed, this time while participants kept running. After this test phase, participants were allowed to rest until the heart rate returned to the baseline level. At this point, a second baseline (T2) was measured. After a short break, the whole procedure was repeated. (B) Paradigm used to measure duration perception after running. This was similar to that described above (A) except the rest period before baseline measurement (T1) was 8 min instead of 3 min, and the test measurements were made after the participant stopped running while the heartrate reverted to baseline.

Figure 2

Heartrate parameters. Average heart rate with ±95% C.I. as a function of time after running onset. In all the experiments (A–C), the heart rate gradually reaches the target value (see methods) within 3 min, then remains stable around that value for the subsequent 5 min of running.

Figure 3

Results on aggregate data for the duration (A,B) and numerosity (C) tasks. Test stimuli magnitudes were plotted against the proportion of “same” responses and fitted with Gaussian functions. The peaks of the fits (arrows) correspond to the PSE (600 ms or 24 dots). A leftward shift (relatively lower peaks values) corresponds to an overestimation of the duration or numerosity of the test stimuli.

Figure 4

Bootstrap Z-test on aggregate data. Distributions of fitted peaks for the duration (A,B) and numerosity (C) tasks. Overlapped distributions indicate no difference between conditions. p-values represent Z-test significance level: *** p < 0.01 > α = 0.017, Bonferroni corrected for three comparisons. n.s.–nonsignificant.

Figure 5

Individual data and perceptual biases magnitudes. Scatter plots showing PSE (log units) separately calculated for each participant (open symbols). Filled symbols refer to group averages. Symbols falling below the equality line (dashed line) reflect lower PSEs in the running condition, hence an overestimation of duration or numerosity. (A) Duration PSEs during running against first baseline (T1). (B) Duration PSEs after running against first baseline (T1). (C) Numerosity PSEs during running against first baseline (T1). (D) Bar plot showing estimation biases relative to the baseline condition (T1), normalized by the reference stimulus (600 ms or 24 dots). Bars represent the three experiments: duration while running, duration after running, and numerosity while running, respectively. Bars are between participant’s average, error bars are ±1 SEM. Individual data are represented by symbols. *** p < 0.001 n.s. not significant.

Figure 6

Duration perception after running. As in Figure 3B, stimuli durations were plotted against the proportion of “same” responses and fitted with Gaussian functions with the peak of the fits (arrows) corresponding to the test duration matched with the reference (600 ms). Blue curves report the aggregate data before running (baseline T1), the red curves the data collected after the running phase. To test whether the effect was detectable in the very first trials after the run phase, the analyses were performed on two sub-set of the data: the first (A) and second half (B).

Figure 7

Perception precision. Bar plot showing estimation precision (Wfs) for duration (A,B) and numerosity (C) tasks. Bars show average, error bars are ± 1 SEM, and circles are individual data. (D) Average Wfs for the duration and numerosity conditions. *** p < 0.001. n.s.–nonsignificant.

Figure 8

Correlation between numerosity and duration Wfs. Scatter plot of individual Wfs as averaged across conditions for duration and numerosity.