NEURAL MECHANISMS OF ROBOT-INDUCED HALLUCINATIONS IN HEALTHY PARTICIPANTS AND SYMPTOMATIC HALLUCINATIONS OF NEUROLOGICAL AND PSYCHIATRIC ORIGIN

Abstract Background The ability to recognize whether sensory consequences have been self-generated or externally produced is an important element of motor control and self-monitoring. Deficits in self-monitoring have been proposed to cause abnormal bodily experiences and psychotic symptoms such as hallucinations. A recent study designed a robotic system that applies sensorimotor stimulation in healthy subjects and safely induces mild presence hallucinations (PH) and passivity experiences (Blanke et al., 2014). PH are defined as the sensation that someone is close by when actually no one is present and passivity experiences are characterized by perceptions or beliefs that an external agent is controlling one’s actions, perceptions, and/or thoughts. Although, both symptoms occur in schizophrenia and Parkinson’s disease, their neural mechanisms are unknown. Methods Here, we first investigated the neural mechanisms of PH and passivity experiences in 25 healthy subjects. We developed a new MR-compatible robotic system able to generate the aforementioned sensorimotor conflicts, while recording subjects’ brain activity using fMRI. In addition, we applied lesion network mapping to 11 neurological patients with symptomatic PH and compared both populations to find a common PH network. In a final step, we investigated the relevance of the PH network by analyzing the connectivity of the PH network in resting state fMRI data from 58 psychotic patients. Results We first evaluated the regions associated with the general sensorimotor conflict, which revealed the left sensorimotor area, the left putamen, the right inferior parietal lobule and the right cerebellum. Then, we analyzed the regions that were more activated during the condition eliciting PH and passivity experiences and found the inferior frontal gyrus (IFG), the insula, the superior medial gyrus and the middle temporal gyrus (MTG). Comparison of these two networks with the symptomatic PH network in neurological patients highlighted the IFG, the MTG and the vPMc. The resting state analysis within those regions in psychotic patients revealed no global differences between the groups but a functional connectivity decrease between MTG and IFG (bilaterally) specific for psychotic patients experiencing passivity experiences. Discussion Collectively, we showed that through the use of a robotic system generating sensorimotor conflicts, the neural correlates of induced-PH and passivity experiences can be studied in healthy subjects in a controlled manner. In addition, we found two networks associated with induced-PH and passivity experience. Of these regions, three were also recruited in patients with PH of neurological origin, forming the PH network. MTG-IFG connectivity in the PH network was altered selectively in psychotic patients with passivity experiences, revealing the relevance of the neural mechanisms of PH and passivity experiences in psychosis.