Tempo rubato : animacy speeds up time in the brain

Abstract

Background: How do we estimate time when watching an action? The idea that events are timed by a centralized clock has recently been called into question in favour of distributed, specialized mechanisms. Here we provide evidence for a critical specialization: animate and inanimate events are separately timed by humans.

Methodology/principal findings: In different experiments, observers were asked to intercept a moving target or to discriminate the duration of a stationary flash while viewing different scenes. Time estimates were systematically shorter in the sessions involving human characters moving in the scene than in those involving inanimate moving characters. Remarkably, the animate/inanimate context also affected randomly intermingled trials which always depicted the same still character.

Conclusions/significance: The existence of distinct time bases for animate and inanimate events might be related to the partial segregation of the neural networks processing these two categories of objects, and could enhance our ability to predict critically timed actions.

Figure 1. Schematics of the interception experiments.

(a) Scene (at 17-m apparent viewing distance) displayed during the static trials of the interception experiments. The ball was thrown from the building and hit ground at the centre of the red cross (magnified for clarity in the Figure). Different positions of the ball during its motion are shown for illustrative purposes only. (b) Single frames from the different types of character animation in the dynamic trials. Here and in the following figures, BM stands for Biological-Motion, UD for Upside-Down, TS for Time-Shifted, RT for Rigid-Translation, DP for Double-Pendulum, and WH for Whirligig. The RT character is depicted in different positions for clarity, and the motion arrow was not present in the actual movie. (c) Time sequence of events during each trial.

Figure 2. Interception timing error (TE) as a function of character type.

Ensemble average TE (± s.e.m.) was computed for all static and dynamic trials of all sessions involving the six characters of Fig. 1b. TE was computed as the difference between the button-press response time and the duration of ball descent. Negative (positive) values correspond to early (late) responses.

Figure 3. Interception timing error (TE) in individual subjects.

Mean TE (± s.e.m.) was computed over all animate characters (Biological-Motion, Upside-Down and Time-Shifted) and over all inanimate characters (Rigid-Translation, Double-Pendulum and Whirligig), and is plotted as An (black) and In (gray), respectively. Left and right bars in each panel correspond to the data for static and dynamic trials, respectively.

Figure 4. Average (± s.e.m.) difference between the mean timing error in dynamic trials and the corresponding value in static trials (Delta TE).

Delta TE was computed over the 7 subjects who performed all 6 experimental sessions.

Figure 5. Response adaptation in consecutive static trials.

Sequences of 3 or more consecutive static trials were extracted from all experiments and subjects. The change of timing error in each consecutive trial of the sequence relative to the first trial is graphed as a function of the serial position i of the corresponding trial (n = 995 trials for i = 1–3, n = 396 for i = 4, n = 177 for i = 5). The first trial of the sequence was preceded by one or more dynamic trial. Notice that the consecutive static trials were not identical, because either ball descent duration or apparent viewing distance varied between any two consecutive trials due to the randomization procedure.

Figure 6. Mean (± s.e.m.) animacy rating computed across all subjects.

Ratings for different characters are color-coded (see right inset), and the values for each of the 9 different semantic pairs are plotted in different columns (bi-polar words on the abscissa). Ratings could vary between 1 and 7, higher ratings denoting greater animacy.

Figure 7. Schematic of the binary choice experiments on duration judgements.

Figure 8. Perceived duration of the standard flash in an animate or inanimate context.

(a–b) Psychometric functions for subject P.C., static (a) and dynamic (b) trials. The graphs show the proportion of times the comparison stimulus appeared to last longer than the standard (360 trials in each panel, 40 repetitions for each of the 9 comparison durations). Data from Biological-Motion and Whirligig sessions are plotted with black and brown symbols, respectively. The vertical lines (placed on the 50% point of the psychometric functions) denote the mean PSEs of the different conditions, and the horizontal error bars denote the 95% confidence intervals of mean PSE. (c) Average values over all subjects of the difference between the PSE for Biological-Motion and that for Whirligig (vertical error bars show the s.e.m.). Negative values indicate that the PSE of Biological-Motion was shorter than that of Whirligig, for both static (red) and dynamic (blue) trials.