Thursday 27 July 2017

How do we judge the speed of human movements?

The speed at which people move - their gestures, walking pace, and so on – conveys useful social information about the meaning and intent behind their actions, as well as clues about their emotional state and temperament. A rapid flick of an eyebrow, for instance, is a common form of greeting while a slow rise and fall can indicate surprise or fear. Walking speed is generally slower when someone is feeling sad rather than happy.

In order for us to use information about action speed in these ways, we need to be able to estimate speed reliably. In this project we asked two questions: How do we perceive the speed of human movements? Does the brain have special neurones that respond according to human movement speed?

In a newly published research article, we attempt to answer these questions by studying how observers make judgements about movement speed in video clips. We report evidence that speed judgments are not fixed and stable, but rely on an internal standard or norm for movement speed that can be altered by relatively short periods of prior experience. The participants in our experiments viewed short video clips of walking or running human figures (recorded on the local high street or at the London Marathon). The playback rate of the clips varied between slow-motion and fast-forward, and participants were asked to judge whether the movement speed of the figures appeared too slow or too fast. This should be an easy task, but we found that participants showed consistent biases in their judgements.

This  clip shows a slow-motion video similar to those we used in the research:



After viewing slow-motion movements for a short period, normal-speed playback appeared to be too fast, and had to be slowed down in order to appear normal. The opposite effect occurred after viewing fast-forward movement for a while. So our judgements of speed are unconsciously influenced by previously viewed speeds. This ‘adaptation’ effect has some interesting implications.

In the silent movie era, cine cameras and projectors were hand-cranked, so playback rate in the movie theatre was very variable (typically between 16 and 24 cine frames per second, depending on how quickly the projectionist cranked the projector, often influenced by programming considerations). Our results indicate that the variation in playback rate was accepted by moviegoers because they quickly adapted to whatever playback rate was used; after a short while the actors appeared to be moving normally, whatever the playback rate.

In a modern context, after you have driven along a motorway at 70mph for a while, you may have had the experience that upon leaving the motorway it is easy to misjudge slow speeds, so your exit speed may be too fast. Our results indicate that this may occur because, perceptually, 70mph became ‘normal’ speed after spending some time on the motorway, causing a speed of 30mph on the slip road to appear slower than it usually does while driving in a city. The neurones in your brain had acclimatised to fast motorway speeds.

Slow-motion video replays from security cameras are increasingly used in criminal prosecutions, to help juries decide whether the crime was premeditated. A recent study in the USA found that after participants had viewed slow-motion replays of criminal acts they were more likely to decide that the defendant had acted in a premeditated way. Intentionality judgements are affected by perceived speed of movement (quick actions tend to be seen as off-the-cuff, without thought). So changes in perceived speed caused by the adaptation effect we report may well be a source of bias influencing legal decisions based on video evidence.

Next season the Football Association will review video footage of fouls in football matches in order to decide whether a player intended to deceive the referee by feigning injury or pretending to have been fouled. If these reviews involve repeated viewing of slow-motion replays, they may well be biased due to the adaptation effect we reported. Offending players may appear to move naturally but have more time, because the total duration of the slow-motion video is longer.

Further reading
Caruso, E. M., Burns, Z. C., Converse, B. A. (2016). Slow motion increases perceived intent. Proceedings of the National Academy of Sciences, 113, 9250–9255.

Mather, G., Sharman, R.J., Parsons, T. (2017) Visual adaptation alters the apparent speed of real-world actions. Scientific Reports, 7, 6738.

Morewedge, C. K., Preston, J., Wegner, D. M. (2007). Timescale bias in the attribution of mind. Journal of Personality and Social Psychology93(1), 1-11.