Search

When interacting with the world, we need to distinguish whether sensory information results from external events or from our own actions. The nervous system most likely draws this distinction by comparing the actual sensory input with an internal prediction about the sensory consequences of one's actions. However, interacting with the world also requires an evaluation of the outcomes of self-action, e.g. in terms of their affective valence. Here we show that subjects' perceived pointing direction does not only depend on predictive and sensory signals related to the performed action itself, but also on the affective valence of the action outcome: subjects perceived their movements as directed towards positive and away from negative outcomes. Our findings suggest that the non-conceptual perception of the sensory consequences of self-action builds on both sensorimotor information related directly to self-action and a post hoc evaluation of the affective action outcome.

We defend our multifactorial weighting model of the sense of agency and our critique of the comparator model ( Synofzik, Vosgerau, & Newen, 2008 ) against the critiques that have been brought forward by Carruthers, 2012 , Wong, 2012 . Building on the specification of our model that emerges from this response, we will suggest a distinct mechanism how weighting of different agency factors might work: internal and external agency cues are constantly weighted according to their reliability in a given situation. Thus, the weighting process underlying the sense of agency might follow the principles of optimal cue integration. We review recent empirical evidence for this hypothesis, demonstrating that the multifactorial weighting model is not only testable, but has in fact already received first empirical support.

The neurocognitive structure of the acting self has recently been widely studied, yet is still perplexing and remains an often confounded issue in cognitive neuroscience, psychopathology and philosophy. We provide a new systematic account of two of its main features, the sense of agency and the sense of ownership, demonstrating that although both features appear as phenomenally uniform, they each in fact are complex crossmodal phenomena of largely heterogeneous functional and (self-)representational levels. These levels can be arranged within a gradually evolving, onto- and phylogenetically plausible framework which proceeds from basic non-conceptual sensorimotor processes to more complex conceptual and meta-representational processes of agency and ownership, respectively. In particular, three fundamental levels of agency and ownership processing have to be distinguished: The level of feeling, thinking and social interaction. This naturalistic account will not only allow to "ground the self in action", but also provide an empirically testable taxonomy for cognitive neuroscience and a new tool for disentangling agency and ownership disturbances in psychopathology (e.g. alien hand, anarchic hand, anosognosia for one's own hemiparesis).

Recent work has demonstrated that the sense of agency is not only determined by efference-copy-based internal predictions and internal comparator mechanisms, but by a large variety of different internal and external cues. The study by Moore and colleagues [Moore, J. W., Wegner, D. M., & Haggard, P. (2009). Modulating the sense of agency with external cues. Conscious and Cognition] aimed to provide further evidence for this view by demonstrating that external agency cues might outweigh or even substitute efferent signals to install a basic registration of self-agency. Although the study contains some critical points that, so we argue, are central to a proper interpretation of the data, it hints at a new perspective on agency: optimal cue integration seems to be the key to a robust sense of agency. We here argue that this framework could allow integrating the findings of Moore and colleagues and other recent agency studies into a comprehensive picture of the sense of agency and its pathological disruptions.