The Role of Velocity in Affect Discrimination

Abstract

The Role of Velocity in Affect Discrimination Helena M. Paterson ([email protected]) Glasgow, G12 8QB Frank E. Pollick ([email protected]) Department of Psychology, 58 Hillhead Street, Glasgow, G12 8QB Anthony J. Sanford ([email protected]) Department of Psychology, 58 Hillhead Street, Glasgow, G12 8QB Abstract Two experiments are described that examine the role of speed in the categorisation of affective biological motion displays. For the first experiment movements were recorded for 10 affects and the point-light animations of them were shown to participants in a recognition task. The resultant confusion matrices were analysed using the ALSCAL multi-dimensional scaling procedure and produced a 2-dimensional psychological space. The psychological space for discrimination was similar to that from recent models of experienced affect in that the first dimension corresponded to the activation dimension from these models. A strong correlation between the movement speed and the activation dimension confirmed the finding. From these results it would appear that the mapping between stimulus properties and representation of activation in affect is a fairly direct one. For the second experiment more sad, angry and neutral movements were collected. New movements of different duration, but identical spatial displacement were made using an interpolation algorithm. Observers viewed the movements as point light displays and heir task was to rate intensity of affect. Results from this experiment indicate speed plays a major role in modulating the intensity of activation in perceived affect. Introduction Humans can easily tell each other apart and interpret subtle differences in behaviour that communicates intentions, identity and emotions easily. Cues from facial features are used for much of this recognition, however, highly impoverished stimuli - such as point- light displays - convey sufficient information for the recognition of such person properties (Barclay, Cutting & Kozlowski, 1978, Cutting & Kozlowski, 1977; Dittrich, Troscianko, Lea & Morgan, 1986; Hill & Pollick, 2000; Kozlowski & Cutting, 1978; Mather & Murdoch, 1994; Runeson & Frykholm, 1981; Runeson & Frykholm, 1983; Walk & Homan, 1984). The cues that convey this information in biological motion are of primary interest to us and using point light displays of human arm movements, we have concentrated on the recognition of emotion from biological motion. In exploring the way in which humans recognise affect, it is possible not only to look at the accuracy with which an affect is recognised, but also at the structure of the representation of affect. A number of models for the structure of experienced affect have been suggested that resemble each other in a number of factors (Russell, 1980; Watson and Tellegen, 1985; Thayer, 1989, Larsen and Diener; 1992; Feldman, Barrett and Russell, 1999). The similarities between these models are that the structure of affect is a two- dimensional and continuous structure. This structure is referred to as a circumplex model (Feldman, Barrett and Russell, 1999). The two dimensions of the circumplex models are bipolar and independent. One dimension represents valence (for instance hedonic tone, pleasant – unpleasant) and the other, arousal or activation (arousal – sleep/ activated – deactivated). The models are also continuous, with affects falling on a circle, centred on the origin of the psychological space defined by the two dimensions. Although these models were established as representing one’s own experience of affect, there is recent evident to suggest that experience and perception interact when observing another person’s actions (Decety and Grezes, 1999; Rizzolatti, Fadiga, Gallese and Foggassi, 1996). Additionally biological motion relies on both specialised bottom-up processes of motion detection (Mather, Radford and West, 1992; Neri, Morrone and Burr, 1998) and interactions between these and top-down processes (Shiffrar and Freyd, 1990, 1993; Thornton, Pinto and Shiffrar, 1998). In the case of affect perception such a top-down process may well originate from the influence of an internal structure of affect. It seems reasonable, therefore, to further explore the possible relationship between perception and structure of affect. There is also, currently, little research that concentrates specifically at the recognition of emotion