“Time flies in the presence of angry faces”… depending on the temporal task used!

Abstract

A number of studies have reported that the perception of an arousing emotional stimulus, such as an angry face, results in temporal overestimations which are probably due to the speeding up of a clock-like system. The aim of the present study was to examine whether this emotional effect can be generalized to all temporal tasks irrespective of the underlying cognitive processes involved in each task. Five different temporal tasks involving the presentation of neutral and angry faces were therefore tested: bisection, generalization, verbal estimation, production and reproduction. Our results showed an overestimation of time for the angry compared to the neutral faces in the temporal bisection, verbal estimation and production tasks but not in the temporal generalization and reproduction tasks. Moreover, the results obtained in the temporal verbal estimation and production tasks suggest that this temporal overestimation of the angry faces was associated with relatively more accurate estimates. The involvement of both arousal and attention mechanisms in the effect of emotional facial expressions on time perception is discussed in the light of the differences in the impact of the same emotional stimulus as a function of the temporal task considered.

Introduction

It is well-known that humans, like other animals, possess neural mechanisms which enable them to estimate time accurately. However, since the pioneering works on temporal distortions (François, 1927, Hoagland, 1933), a number of empirical findings have made it clear that internal and external factors affect these temporal mechanisms. These factors include emotions, which change each individual's sense of time and make it fly past quickly or drag slowly.

Over the last decade, the question of how emotions affect our time perception has been the focus of renewed interest (for a review, see Droit-Volet and Gil, 2009, Droit-Volet and Meck, 2007, Stetson et al., 2007, Wittman and van Wassenhove, 2009). This is due to the recent introduction of standardized and validated emotional materials which make the data more reliable and deserving of further scientific consideration (Coan & Allen, 2007). For example, time distortions have been examined using the International Affective Picture System (IAPS, Lang, Bradley, & Cuthbert, 2005) by Angrilli, Cherubini, Pavese, and Manfredini (1997), while Noulhiane, Mella, Samson, Ragot, and Pouthas (2007) employed the International Affective Digitalized Sounds System (IADS, Bradley & Lang, 1999). Other researchers have tested the impact of emotional facial expressions on time perception (e.g. Doi and Shinohara, 2009, Droit-Volet et al., 2004, Gil et al., 2007, Tipples, 2008). It is, however, difficult to compare the reported findings because time distortions vary as a function of the presented emotional stimuli, with each stimulus having specific characteristics with regard to adaptation and timing. For example, arousing stimuli which elicit disgust, such as disgusting pictures or faces expressing disgust, do not all have the same effect on time, with the former producing a subjective lengthening effect (Angrilli et al., 1997) and the latter not producing any significant effect (Droit-Volet and Gil, 2009, Gil and Droit-Volet, in press). Building on the ideas put forward by the pioneers of studies on emotions (Darwin, 1998, Dewey, 1894, Mead, 1934), it is clear that the faces observed in everyday life exhibit a high degree of specificity which enables individuals both to collect important information about the dynamic complexity of the social context and to take the appropriate action (e.g. Fridlund, 1994, Frijda and Tcherkassof, 1997, Keltner and Haidt, 2001). The aim of the present study was therefore to investigate the effect of facial expressions on the judgment of time in greater detail.

Our team recently undertook a series of studies to test the effect of emotional faces on time perception by means of a temporal bisection task which has already been extensively used both with animals (e.g. Church and Deluty, 1977, Meck, 1983, Orduna et al., 2007) and with human adults (e.g. Allan and Gibbon, 1991, Droit-Volet et al., 2007, Wearden and Ferrara, 1995, Wearden et al., 2006). In these studies, the participants were initially exposed to a short (e.g. 400 ms) and a long standard duration (e.g. 1600 ms) in the form of a displayed oval. They were then exposed to different comparison durations that were either similar to or lay between the standard durations. These were presented in the form of neutral faces and faces expressing emotion. The participants were instructed to categorize the presentation durations of the faces as being more similar to the short or to the long standard duration. The findings showed that the faces expressing anger were systematically judged to be presented for longer than the neutral faces, thus shifting the bisection function toward the left and lowering the bisection point (BP, also known as the point of subjective equality). This subjective lengthening effect produced by the perception of angry faces has since been replicated in other studies which have used the bisection task in both adults (Doi and Shinohara, 2009, Droit-Volet et al., 2004, Effron et al., 2006, Mondillon et al., 2007, Tipples, 2008) and children (Droit-Volet and Gil, 2009, Gil and Droit-Volet, in press, Gil et al., 2007). This effect is explained by the fact that the perception of angry faces is arousing. This prepares the body for action and therefore accelerates the biological clock mechanism. When angry faces are perceived, the internal clock is therefore thought to generate more biological units (pulses, oscillators), with the result that time is judged to be longer.

In studies of time perception, researchers often arbitrarily choose a temporal procedure without performing any in-depth examination of the differences in the cognitive processes which underpin temporal judgments in different temporal tasks. This is primarily due to empirical findings which have revealed high levels of correlation between the results of different temporal tasks (Wearden, 2003, Wearden and Lejeune, 2008). However, as suggested by models of temporal information processing (Scalar Expectancy Theory, SET; Gibbon et al., 1984, Harrington et al., 2004, Rubia and Smith, 2004), subjective time results from an interaction between different cognitive processes (i.e. clock-related, memory and decisional processes) each of which may be involved to a greater or lesser degree as a function of the employed temporal task and therefore have a different effect on the time judgment that is ultimately produced. For instance, Baudouin, Vanneste, Isingrini, and Pouthas (2006) suggested that variations in the production of duration are primarily dependent on the internal clock rate, whereas the reproduction of duration is dependent on working memory capacities. To date, the effect of emotional faces on time judgment has only been tested using the temporal bisection task. The question that is raised here is whether the lengthening effect obtained in response to an arousing emotional expression – angry faces – occurs in all temporal tasks irrespective of the type of temporal judgment involved. In general, it would seem that temporal judgments made in response to arousing emotions are simply the by-product of an acceleration of the internal clock. If this is indeed the case, we should expect the angry face-related effect to occur in all tasks. An alternative hypothesis is that the temporal judgment of emotional faces depends on an interaction between distinct cognitive processes which differ as a function of context and temporal task. If this is the case, the effect produced by emotional faces should vary as a function of the temporal task.

The aim of the present study was thus to examine the judgment of the presentation duration of angry faces compared to that of neutral faces in different temporal tasks. We tested two temporal discrimination tasks: a temporal bisection task and a temporal generalization task. In the temporal bisection task, the subjects had to judge each comparison duration as short or long, while in the temporal generalization task they had to compare a presented duration (longer than, shorter than, or equal to) with a standard duration stored in memory. However, these tasks do not give us any direct insight into the extent to which subjective time is distorted compared to real time. Three additional temporal tasks were therefore tested: verbal estimation, temporal production and temporal reproduction. In the verbal estimation task, the participants had to assign a verbal label to a stimulus duration. In the production task, they had to produce a target duration corresponding to a verbally presented indication. More precisely, they had to press a key to stop a stimulus when they thought that its duration corresponded to the indicated value. In the reproduction task, the subjects had to reproduce the duration of a stimulus. They were presented with a target stimulus duration and were then presented with a stimulus which they had to stop when they thought that its duration corresponded to the previously perceived duration. In all these temporal tasks, the participants were presented with either angry or neutral faces.