M58. EVALUATING WEIGHTINGS OF PREDICTIVE AND POSTDICTIVE PROCESSES IN ABERRANT SENSE OF AGENCY IN SCHIZOPHRENIA

Abstract Background Sense of agency (SoA) refers to the subjective feeling that one is the agent who causes and controls one’s own actions and those effects on the external world. It’s widely known that patients with schizophrenia show aberrant SoA, but its cognitive and neural mechanisms remain unclear. Previous studies suggest that abnormal SoA is probably due to impaired predictive processes, such as predicting sensory feedbacks according to one’s own motor intention or motor commands. However, dysfunction of the prediction process alone cannot explain various symptoms of schizophrenia. The impairment in postdictive processes in patients with schizophrenia is also considered to be important to explain their symptoms. However, how the predictive and postdictive processes actually contribute to the judgment of agency in patients with schizophrenia remains unknown. In the present study, we focus on the functional contribution of the two processes at sensorimotor level to the judgment of agency at the cognitive level, rather than the performance of these functions. Our approach allows us to evaluate the weightings of each contribution quantitatively, and to associate such weightings to symptoms for each individual. The results of our study point to potential useful diagnosis, and can provide important knowledge for treatment in future research. Methods We use two sensorimotor tasks (i.e. the reaching task and the control detection task) and one explicit agency judgment task in this study. In all tasks, participants move one or several dots on a screen by moving a mouse. The speed of the dots’ motion always corresponds to participants’ mouse movement, but the direction of the dots’ motion is combined with both participants’ motion and prerecorded motion, and is applied with a certain angular bias. In the reaching task, participants move the dot to touch a destination on screen as quickly as possible. This task is designed to measure the extent of prediction errors at the sensorimotor level, which is considered as an important signal for the predictive process underlying agency judgment. In the control detection task, participants move three dots simultaneously, and identified one dot that can be partially controlled by their mouse movement, while the other two dots move in pre-recorded directions. This task does not require explicit judgment of agency. It measures the postdictive process that to detect the regularity between the visual feedbacks and one’s own movements at the sensorimotor level. At last, in the explicit agency judgment task, participants made binary judgment (i.e. yes or no) regarding whether they feel that they controlled the moving direction of the dot or not in each condition. In summary, the two sensorimotor tasks measure the output of predictive and postdictive processes at the sensorimotor level, respectively, and the explicit agency judgment task measures agency judgment at the cognitive level. We designed six experimental conditions, in which the actual angular error and regularity were systematically manipulate. Such manipulation creates variability in the performance of each task, and allows us to use structure equation modelling to analyze the contributions of the performance in the two sensorimotor tasks to the explicit agency judgment task, and to compare the results with healthy controls. We focus on the functional structure of the sensorimotor sub-component and the cognitive judgment of SoA. Results The data collection is currently in progress. We plan to finish it by February. Discussion We predict weakened contributions from the two sub-components to SoA in patients, compared to healthy control. We plan to link the quantitatively estimated weightings to individual symptoms including both positive and negative symptoms.