Most patients with Parkinson's disease (PD) are not only affected by motor symptoms (e.g., tremor, rigidity, bradykinesia), but also experience hallucinations. Importantly, hallucinations are a risk factor for a more rapid cognitive decline, dementia, and are associated with higher mortality. Notably, a group of hallucinations, called minor hallucinations, occurs at an early stage of the disease and is an indicator of a more severe progression of PD. However, despite the high prevalence and clinical relevance the neural mechanisms of hallucinations remain poorly understood, in large part due to the absence of procedures and methods to induce and quantify hallucinations. Recently, by combining knowledge from robotics, cognitive neuroscience, and clinical research, our laboratory has developed a robotic system enabling the induction of a specific minor hallucination - presence hallucination (PH) - by experimentally exposing participants to different sensorimotor conflicts (robot-induced PH). This thesis is aimed at better understanding the underlying mechanisms of robot-induced PH as well as symptomatic-clinical PH, by focusing on their animate-social nature. Specifically, I aimed to quantify PH and assess its behavioral, neuropsychological and neural correlates in healthy subjects and PD patients by combining immersive virtual reality (VR), robotics, digital technologies, psychophysics and neuroimaging. First, I developed an implicit measure of robot-induced PH through a novel social perceptual task involving a VR-robotics method where participants are asked, after PH inducing sensorimotor stimulation, to estimate the number of people they perceived in a room briefly seen in a VR scenario. With this method, I established in healthy subjects that elevated numerosity estimation of humans in VR is a digital marker for experimentally induced PH. Second, I adapted and extended this method to PD patients, in an online study, where I have shown that elevated numerosity estimation of humans is also a marker for symptomatic hallucinations in PD. Crucially, in both projects, the overestimation effect was specific to animate-social stimuli (humans) and was absent for inanimate control objects, revealing a social perceptual deficit underlying both robot-induced PH and symptomatic-clinical hallucinations in PD. Third, using EEG, I investigated the neural correlates of robot-induced PH in healthy subjects during our social perceptual task and demonstrated that elevated numerosity estimation through robot-induced PH modulates parietal activity at a specific time window (P2p component), which was previously associated with (non-social) numerosity estimation. Finally, using behavioral analysis and fMRI, I investigated social perceptual deficits in PD, linking hallucinations to functional connectivity impairments in structures commonly involved in social visual perception and numerosity. In summary, this thesis made a two-fold advancement: the development of an innovative method of measuring clinical symptoms of PD hallucinations and a newly established link between behavioral and neural correlates of PH and social perception alterations. In line with this, the present thesis suggests a reconsideration of PD as a movement disorder to a movement-and-social perceptual disorder, for better understanding of the nature of patients' deficits and development of more effective future diagnostic and treatment targets.