Column

(No.14) Perception of Past Time

Research on human perception has a history of more than 150 years. Newly introduced MRI (magnetic resonance imaging) has re-ignited the research with many new discoveries. Historically, E. H. Weber and G. T. Fechner's contributed "Strength of perception is proportional to the logarithm of stimulant strength," around 1860. This is still a landmark contribution. Since then, research on psychological experiments have become popular and yielded many important results. Two examples are the establishment of three primary colors and color space. However, very little research on time perception has been done because of experimental difficulties. Here, I discuss memory regarding time perception. According to Weber-Fechner's law, outside stimulant s with additional stimulant causes a perceptional signal increase from i to , which is expressed as


Here, k is a constant. Integrating both sides of the above equation,

k ยท log s = i

This expression says that the perceived signal is proportional to the logarithm of the stimulant. The most well-known example is brightness i caused by light beam energy s.

Is this relation valid for time perception?

What is the human's time perception mechanism? Fujii et al. (N. Fujii and A. M. Graybiel, "Representation of Action Sequence Boundaries by Macaque Prefrontal Cortical Neurons," Science, pp. 1246-49, Aug. 29, 2003) found that groups of pulse trains were observed corresponding to event boundaries. This may be the cause of time perception. As the stored events are memory, time perception may be closely related to memory. Extrapolating known mechanisms, the following hypothesis holds:

(Hypothesis) The duration of past event has perceptional width , when the elapsed time was t.

In other words, a one day event that happened 10 days ago has a similar time width in memory of a one year event that happened 10 years ago. This concept is the same as Weber-Fechner's law, which can appropriately be represented along a logarithmic scale (Figure 1).

The event may happen evenly without regard to the time scale. If we display the events on a logarithmic scale, events of long ago have narrower space. To solve this problem, only important events are shown. Less important events are ignored or shown with smaller fonts. This is reasonable, is it not? But someone needs to assign an importance level for each event. This method of display on an uneven time scale event has been used for a long time, especially for historical timelines. As we have less knowledge about the ancient past, this scaling method was reasonable. We should not, however, mix historical events and personal memory on the same level.

Fig. 1 An example of a Calendar:Retrieved events based on personal memory
Fig. 1 An example of a Calendar:Retrieved events based on personal memory

Digital memory allows more and more storage at less and less cost. Now people want easy retrieval rather than more storage. We can use keywords to retrieve text data, but there is no effective way to retrieve photographs, music, or video data, which are commonplace in the modern family. In the example in Figure 2, thumbnail sizes are proportional to the blank period of the time before it was taken. After longer blank time, the next event may be remembered more strongly.

The new approach has just emerged. Further progress is expected in achieving easy user interfaces.

Fig. 2
Fig. 2

(Ej, 2004.08)

Back to index

Page Top