B. Selection of low, moderate and high amounts of information

For the needs of the subsequent experiments (social and non-social motor and superordinate exp.), 3 different amounts of visual information (corresponding to three different movie segments) were selected among the 12 controlled ones. To do so, a Weibull psychometric function was fit with the restricted least squares method to the data (we used a Weibull function because such function generally provides a good model for contrast discrimination and detection data ; see Nachmias, 1981; Wichmann & Hill, 2001). We then computed the amount of visual information corresponding to the inflexion point of the sigmoid curve within each experiment (shown as a blue dot on the resulting regression curves, see figure 10). The duration corresponding to the inflexion point(1560ms –moderate) and the two nearest durations (1480 –, low and 1640 ms – high) were selected for the covert blocks. This choice ensured equal timing steps as well as similar performance differences between the three levels. The 1880 ms duration (≈100% hit rate) was chosen for the ‘overt’ trials.

The Weibull function is described by:

Fig. 10 . Motor experiment: psychometric curve (green) fit to the cumulative distribution of participant’s correct responses (red dots) across the 12 movies segments. Responses for the different actions were pooled across. The blue dot refers to theinflexion point of the sigmoid curve. In each experiment, the inflexion point occurs at the following duration: A. Motor: 1576 ms. B. Superord.: 1558 ms. C. Social motor: 1546 ms. D. Social superord.: 1550 ms.