Surprise! You Just Distorted Time.

"Where is it, this present? It has melted in our grasp, fled ere we could touch it, gone in the instant of becoming." - William James, 1890

Extremely dangerous, traumatic, or surprising moments are often accompanied by reports that time seemed to "slow down" or "fly by." The perceptual basis of these subjective temporal distortions is unclear, but not for lack of trying: one recent experiment went so far as to drop subjects off a 400 foot tower while testing their ability to decipher a rapidly flashing string of numbers - a test of perceptual processing speed. Unfortunately, it didn't work. Subjects were no better at deciphering these numbers than they were under more mundane circumstances.

Ironically, temporal distortion may be more noticeable in such mundane experiences. A 2004 study by Tse and colleagues reviews evidence that durations are estimated as somewhat longer as more complex stimuli are being presented - as though time is subjectively expanding. In contrast, when subjects must actively attend to those stimuli or perform a secondary task while estimating durations, they tend to estimate those elapsed durations as slightly shorter - as though time is subjectively contracting.

Tse et al. suggested that these effects might emerge in the same way distraction might impair your ability to count the "ticks" from a clock's second hand. According to this "missed temporal cues" hypothesis, duration judgments are accomplished by attending to "temporal units." When attention is fully directed towards these temporal judgments, fewer "temporal units" are missed and duration estimates increase; however, when attention is divided, more of those "temporal units" are missed and so the estimates decrease.

Tse et al also suggest an alternative hypothesis: "attentional boost." According to the "attentional boost" theory, the processing of relatively low probability stimuli may somehow speed information processing of that stimulus, causing more "temporal units" to be counted.

To distinguish between these hypotheses, Tse et al. combined several methods of temporal estimation with the oddball paradigm, in which relatively low-probability stimuli are embedded inside series of more high-frequency stimuli. (For example, in auditory oddball, subjects might hear a string of sounds like "ROOF ROOF ROOF WOOF ROOF." Bizarre, I know.)

In a first experiment, subjects reliably estimated that an oddball visual stimulus was similar in duration to a much more frequent visual stimulus when, in fact, its duration was around half as long (675 msec vs. 1050 msec)!

A second experiment demonstrated that this temporal expansion of "oddballs" occurs only when the more frequent standard stimuli are longer than ~150 msec in duration. In fact, there's a subjective temporal contraction when standard stimuli are around 75 msec in duration - as though attention cannot be allocated to the oddball quickly enough, and the subsequent blank interval or standard stimulus itself undergoes the subjective temporal expansion, instead of the oddball.

An interesting feature of the data is a peculiar "dip" in the degree of temporal distortion measured by these methods when the more frequent standard stimuli last only around 375 msec. Shorter or longer standards cause more temporal distortions. Furthermore, this dip occurs at different times between individuals, and Tse et al. argue that it could reflect what many consider to be the dual mechanisms of attentional reorienting: a transient component and a sustained component.

Subsequent experiments replicated the effect with a variety of stimuli and temporal estimation methods. These temporal distortions (and to some extent the pecular "dip") were apparent when using auditory stimuli and were present with visual stimuli regardless of whether subjects were asked to rate the magnitude of the oddball's duration, to reproduce the duration of the oddball, or to rate whether the oddball lasted longer or shorter than the average of all previous stimuli (all of these are established psychophysics techniques for duration perception).

Tse et al note their results could be an example of the "time order error" (TOE) in which the second of two sequentially judged stimuli is altered along the dimension of judgment. For example, when two weights of large mass are lifted in succession, subjects tend to say the second is heavier. In contrast, when two weights of small mass are lifted one after the other, subjects tend to rate the second as lighter. So, in the case of temporal distortion, subjective duration lengthens if the oddball is presented for longer than 150 msec, but shortens if the oddball is presented for less than 150 msec. They discuss several reasons to doubt this possibility, including that TOE theory seems to contradict several findings in the duration psychophysics literature, as though it might not apply to duration judgments.

There are clear evolutionary advantages to subjective temporal distortion - as Tse et al note, increasing temporal resolution could allow for greater depth of processing and thus more adaptive responding when an organism is endangered or surprised. Although intuitively far-fetched, visual psychophysics has demonstrated that spatial attention can alter the spatial resolution of vision. Thus it seems at least plausible that "temporal attention" could alter the temporal resolution of duration judgments, causing them to contract or expand. Finally, an interesting question for future research is how these phenomena might interact, given the time-frequency tradeoffs that are inherent to signal processing. For example, increased spatial attention could cause decreased resolution in temporal judgments, and vice versa.

Related Posts:
Reversing Time: Temporal Illusions (this post covers work by Eagleman and colleagues that reveals a perceptual form of latency or "delay compensation)
Reichardt Detectors and Illusory Motion Reversal (This post covers yet other work by Eagleman et al., in which they demonstrate that illusory motion reversal may occur because of "temporally tuned" motion detectors in the visual system)
Tuned and Pruned: Syneasthesia