Keywords

1 Introduction

A critical aspect of human-computer interaction is that individuals must be able to precisely determine when to execute actions to achieve a certain goal. For example, in virtual games, players often need to decide when to move or fight with enemies so that they could click a mouse or press keys timely; in human-robot interaction, users need to accurately estimate when to send an instruction to robot according to some feedback information. Therefore, it is essential to explore time perception during a dynamic human-computer interaction and to investigate the possible influential factors.

The duration of visual stimuli can sometimes be perceived longer or shorter than its actual duration. Previous studies showed that many visual features can distort subjective duration, including stimulus size, brightness, number, complexity, and spatial frequency [1, 2]. Furthermore, static images with implied motion or body movement can affect subjective time [3]. Besides the features of visual stimuli per se, recent evidence suggests that our perception of time also depends on whether we are moving or not, which turns out an important external information interacting with visual sensation, especially when dealing with information on computer/machine interface in a state of motion [4]. Moreover, moving state can lead to compressed or expanded effect of duration perception, and influence stability or accuracy of time prediction [5].

Taken together, some properties of body movements would influence subjective time perception, including motion speed, motion duration, motion direction, action stage and some individual factors during motion. We summarized the relevant experiments and results in Table 1.

Table 1. Properties of motion that influence time perception and the related experiments
Full size table

The neural mechanism of body movements’ influence on time perception might stem from the tight relationship between time and space representation in our brain [12]. Growing number of studies revealed that motoric brain structures may form the core component of a neural network supporting a wide range of timed behaviors. Supplementary motor area (SMA) is not only a part of motoric brain structures, but also plays a key role in time processing as part of the striato-cortical pathway verified by animal studies, human neuropsychology and neuroimaging studies [13].

In human-computer interaction, our time perception of visual stimuli often involves durations lasting below or above a second, namely the sub-second or supra-second timing scale. A large body of neuroimaging studies suggested that distinct neural systems are recruited for the measuring of sub-second and supra-second intervals [14,15,16,17]. For example, Pouthas et al. [18] found that the activation of some brain regions increases with increasing time, which demonstrated that sub-second and supra-second might be based on different internal clocks. Furthermore, previous studies pointed out that sub-second durations are processed in the motor system, whereas supra-second durations are processed in the parietal cortex by utilizing the ability of attention and working memory to keep track of time [19]. However, no conclusive theories have been presented to explain how body movements influence time perception of different timing scales.

Hand movements are frequently seen when we operate computers. When we play games or immerse ourselves in VR, we often use hands to manipulate or interact with some equipment. The present study aimed to verify how hand movements influence time perception of visual stimuli; and to probe whether there is a difference in duration perception between sub-second and supra-second stimuli and how hand movements modulate them.

2 Method

2.1 Participants

A total of 11 students (3 males) with a mean age of 20.55 years old (range 18–27) participated in the experiment. All participants were right-handed, with normal or corrected-to-normal vision, had no known abnormalities of their motor systems. They gave written informed consent prior to the experiment and were paid for their participation.

2.2 Stimuli and Design

Visual stimuli in the experiment were generated by a computer and presented on a computer screen. The viewing distance to the monitor was about 50 cm. Participants sat at a table and stared horizontally at the center of the screen. Their left hands placed in an opaque carton on the left of the keyboard on the table, and they used right fingers to press keys to respond.

The experiment was divided into a hand movement part and a static part, and the order of them was balanced among participants. The study adopted the method of constant stimuli belonging to the psychophysical method. In the movement part, participants put their left hands in the opaque carton and moved in an anticlockwise circle at a constant speed. In the static part, they rested their left hands in the carton.

The standard visual stimulus was a blue circle which lasted 700 ms at the sub-second scale, and 1400 ms at the supra-second scale. The two different standard visual stimuli scales were presented randomly in both experiment parts, in case of the practice effect. Each standard stimulus was corresponding to five probe durations (Table 2). Each probe stimulus presented ten times in both parts randomly.

Table 2. Standard and probe stimuli durations in the test phase.
Full size table

Before the test phase, there was a training phase which also contained movement and static parts. The duration of standard visual stimulus was 800 ms at the sub-second scale and 1500 ms at the supra-second scale in the training phase (Table 3). Each probe stimulus presented once in both parts. Participants would get their accuracy feedbacks in training phase, while in the test phase, there was no feedback.

Table 3. Standard and probe stimuli durations in the training phase.
Full size table

2.3 Procedure

In each trial, a fixation appeared firstly at the center of the screen in a uniform gray field, and a blue circle as a standard stimulus appeared for a while, followed by a masking stimulus; next, an identical circle as the previous one appeared for some time. Participants need to make a judgment with their right fingers on which circle lasted longer by pressing “1” or “2” keys on numeric keypad. A timeline chart of the experiment is shown in Fig. 1, it illustrated the procedure of the movement part. And the procedure of static part is similar except that participants’ left hands were static for the whole experiment.

Fig. 1.
figure 1

Schematic illustration of the experimental procedure of hand movement part.

Full size image

2.4 Data Analysis

The main dependent variables were the proportion participants judged the probe stimulus longer than the standard stimulus for each scale, as well as the point of subjective equality (PSE: the 50% point of the cumulative logistic distribution function fitted to the psychometric function by the maximum likelihood method to estimate the 50% correct point) for each standard stimulus. Logistic regression was used to relate the percentage of ‘longer’ judgment responses to overall stimulus duration in each condition for each participant. We also calculated the Weber ratio to analyze the temporal precision or sensitivity to the visual stimuli. This ratio is obtained by dividing the difference limen (half of the difference between the duration giving rise to 75 and 25% longer responses of probe stimuli) by the PSE.

3 Results

3.1 Proportion of Longer Probe Durations

For the proportion of probe stimuli being responded as longer ones, a two-way repeated-measures ANOVA was used to analyze the factors of movement states (motional vs. static) and timing scales (sub-second vs. supra-second). The index reflects a tendency of whether the duration perception was expanded or compressed (The main results were depicted in Fig. 2). Results showed that the main effect of movement state was significant (F = 11.1, p = 0.01, \( \eta_{p}^{2} = 0. 5 8 2 \)), the main effect of timing scales was significant (F = 181, p < 0.001, \( \eta_{p}^{2} = 0. 9 5 8 \)), and the interaction effect of the two factors was also significant (F = 12.8, p = 0.007, \( \eta_{p}^{2} = 0. 6 1 5 \)). The post-hoc tests showed that the proportion of longer times of probe stimuli of sub-second was smaller than supra-second, which implied a compressed effect for the perception of sub-second stimuli than the supra-second stimuli.

Fig. 2.
figure 2

Proportion of longer responses of probe duration plotted for each standard stimulus in movement and static condition (left: sub-second scale; right: supra-second scale).

Full size image

The simple effect demonstrated that how hand movement could modulate the distortion of two timing scales, for sub-second, hand movements enhanced the compressed effect as compared to a stationary state (t = −4.89, p = 0.005), while for supra-second, there was no significant difference between movement state and stationary state (t = 0.17, p = 0.998).

3.2 Point of Subjective Equality (PSE)

The responses for each individual were modeled by fitting cumulative Gaussians and were used to calculate the PSE value of each timing scale. The PSE results can further examine the compressed effect for sub-second and the expanded effect for supra-second time perception (sub-second: PSE = 820 ms > 700 ms; supra-second: PSE = 1000 ms < 1400 ms). In addition, a two-way repeated-measures ANOVA was used to analyze the factors of movement states (motional vs. static) and timing scales (sub-second vs. supra-second) on PSE. Results showed that the main effect of movement state did not reach significant (F = 0.18, p = 0.68, \( \eta_{p}^{2} = 0.0 1 8 \)), the main effect of timing scales was significant (F = 1175.45, p < 0.001, \( \eta_{p}^{2} = 0. 9 9 2 \)). However, because the standard stimuli durations were actually different between sub-second and supra-second, the statistical analysis had no actual meaning about the main effect of timing scale. And the interaction effect of the two factors did not reach significant (F = 0.09, p = 0.77, \( \eta_{p}^{2} = 0.00 9 \)).

3.3 Weber Ratio

As for the Weber ratio results, the two-way repeated-measures ANOVA showed that the main effect of movement state did not reach significant (F = 0.001, p = 0.99, \( \eta_{p}^{2} = 0 \)). The main effect of timing scales was significant (F = 37.95, p < 0.001, \( \eta_{p}^{2} = 0. 7 9 1 \)), which indicated that participants were more sensitive to sub-second duration. And the interaction effect of the two factors did not reach significant, which showed that there was no obvious modulation of hand movements on temporal perception precision in different timing scales (F = 0.523, p = 0.486, \( \eta_{p}^{2} = 0.0 5 \)) (The main results were depicted in Fig. 3).

Fig. 3.
figure 3

Weber ratio plotted for sub- and supra-second scales in movement and static condition.

Full size image

4 General Discussion

The present study examined whether hand movements influence perceived presentation duration of visual stimuli in sub- and supra-seconds. We used the blue circle as the visual stimuli presented on screen and asked participants to judge whether the standard circle or probe circle lasted longer, when their left hands moved in an anticlockwise circle at a constant speed or just rested on table.

As a general result, we found that individuals perceived sub-second visual stimuli shorter than the actual duration, which could be viewed as the compressed effect. While for supra-second visual stimuli, individuals perceived them longer than actual duration, namely the expanded effect. And the temporal precision or sensitivity was much higher when perceiving sub-second visual stimuli. In addition, hand movements enhanced the compressed effect of sub-second visual stimuli.

4.1 Different Timing Scales

The bulk of the analyses we performed on our data set yielded results consistent with the previous studies that for all participants, independent with the movement condition, two different “clocks” for evaluating time below and above a second, respectively, were applied in time perception of visual stimuli [20]. We supported the idea from two scalar properties: (1) the results from proportion of longer probe durations and PSE revealed different duration distortions of two timing scales (a compressed effect for sub-second duration and an expanded effect for supra-second duration); (2) the results from Weber ratio suggested different perceiving precision of two timing scales (the precision of sub-second was higher than supra-second).

However, there is no uniform conclusion about which time point is the cut-off point of different duration processing mechanisms. For example, Michon proposed that below 1/2 s, temporal information processing has the attribute of perception processing [21]. While Pöppel emphasized that below 2–3 s, temporal processing should be viewed as time perception, and above 2–3 s, temporal processing should be viewed as time estimation [22]. Therefore, more research is needed to address the question.

4.2 Modulation of Hand Movement on Time Perception

Another main finding of the current study was that hand movements could modulate the time perception of visual stimuli, especially for sub-second scale. Compared with static condition, the compressed effect was enhanced by hand movement condition. According to previous neuroimaging studies, different brain areas were associated with the ability to discriminate sub-second and supra-second intervals [19]. The estimation of the sub-second range tends to recruit the primary sensorimotor and supplementary motor cortices and the cerebellum, which is more like “Automatic timing”. Whereas supra-second timing tends to recruit the posterior parietal, which is associated with “cognitive timing” [17]. Therefore, we can speculate that hand movements involve more activation of primary sensorimotor areas and might lead to the modulation of sub-second estimation.

4.3 The Limitation of the Study

A major limitation in our study is that only one temporal duration in each range of time (sub- and supra-second time) has been considered. In future, we can try to apply more different time durations of the two timing scales, just like in Chen and Cesari [11] ’s study.

5 Conclusion

In summary, although the distortion of sub-second and supra-second is different when individuals perceive a duration of visual stimuli, hand movements can modulate the distortion and enhanced the compressed effect of sub-second visual stimuli. The results further indicate that the processing mechanisms of the two timing scales are distinct. The findings have implications for time perception mechanisms during dynamic human-computer interaction involving body movements.