Hierarchical Control and Sense of Agency: Differential Effects of Control on Implicit and Explicit measures of Agency

Hierarchical Control and Sense of Agency: Differential Effects of Control on Implicit and Explicit measures of Agency Devpriya Kumar (devpriyak@cbcs.ac.in) Centre of Behavioural and Cognitive Sciences, University of Allahabad Allahabad 211002, India Narayanan Srinivasan (nsrini@cbcs.ac.in) Centre of Behavioural and Cognitive Sciences, University of Allahabad Allahabad 211002, India Abstract While recent studies show dissociation between the implicit and explicit aspects of ‘sense of agency’, the mechanisms underlying these different aspects of agency are not yet clearly understood. We argue that the control achieved at different levels of hierarchy is important for different aspects of agency. In the current study, we investigate how changes in control at the perceptual-motor level and at goal level influence implicit and explicit measures of sense of agency. In a given trial, participants were first required to aim at a target in a noisy environment and then shoot at the target. After certain interval, a circle flashed at the location where participant aimed while pressing the trigger. Participants estimated the interval between action and presentation of the circle that acted as a measure of intentional binding, an implicit measure of agency and also rated an explicit sense of authorship. The results suggest that different aspects of agency and dissociation between implicit and explicit aspects of agency are mediated by control achieved at various levels. Keywords: Sense of agency; event-control approach; intentional binding; control; hierarchical system; interval estimation. Introduction With every action that we perform, we not only influence our environment but also modify our conscious mental state of being the agent of the action. This feeling is known as sense of agency (Pacherie, 2011). While elusive, this sense of agency (SoA) is central to our conscious experience and has recently gained popularity among philosophers as well as scientists (see Gallagher, 2006, for a review and insight on related issues). Sense of agency is a complex, multifaceted, phenomenon (Pacherie, 2011). In general two aspects of SoA have been emphasized and studied in detail. Firstly, sense of Agency as first order experience, in which agency is generally linked to the intentional aspect of task. Here an action is considered to be self-generated when the effect of an action matches the intention of the participant (Moore, Lagando, Deal, & Haggard, 2009; Farrer & Frith, 2002), or when agency is linked to bodily movement as in the famous rubber hand illusion (Farrer, Frank, Georgieff, Frith, Decety, & Jeannerod, 2003). This aspect of agency is also called pre-reflective or implicit sense of agency. Secondly, sense of agency as reflective attribution (or sense of authorship), in which participant is asked to report his/her subjective sense of belief in causing an action (Ebert & Wegner, 2010; Haggard & Moore, 2010). Recent models of ‘self’ (Synofzik, Vosgerau, & Newen, 2008) take into account these two aspects of agency. An important measure of agency that has gained prominence in the last decade is intentional binding. Intentional binding refers to the finding that participants perceive the self-generated action and its effect to be temporally closer to each other (Haggard, Clark, & Kalogeras, 2002b). The concept of intentional binding has been linked in the literature strongly to the sense of agency, that is, the experience of agency is greater when intentional binding is stronger. A recent review (Moore & Obhi, 2012) suggests that intentional binding has been associated with implicit measures of agency like efference, sensory feedback, causal feedback, and intentionality. Haggard and Clark (2003) have suggested that when motor cortex is stimulated to produce a movement similar to a voluntary movement, intentional binding is not affected, suggesting that intentional binding does not depend on the sensory signals produced during movement, but rather it depends on the efference copy generated during action planning. Desantis and colleagues (Desantis, Cedric, & Waszack, 2011) showed that, when participant believes that he/she has control over the environment, intentional binding is stronger. In the original study showing intentional binding, (Haggard, Clark, & Kalogeras, 2002b) the outcome of a participant generated action was manipulated in terms of intention (intended or unintended). Results indicate that intentional binding is stronger for the intended effect compared to the unintended effect. These and many more studies indicate a strong link between the implicit measures of agency and intentional binding, suggesting that intentional binding can be used as a reliable measure of implicit sense of agency. In addition, researchers have investigated the relationship between intentional binding and explicit sense of agency or reflective sense of agency (Moore & Obhi, 2012). In one such study (Ebert & Wegner, 2010), participants were presented with a picture, which can either move congruent or incongruent to the direction participant moved the joystick. The delay between the participant’s movement of the joystick and the movement of the object on the screen