Deep brain stimulation and psychosis: A case series and two candidate causal brain circuits

Schizophrenia and psychosis are debilitating conditions with suboptimal treatment options. Deep brain stimulation (DBS) offers promise, but effective treatment targets remain undefined. Examining cases in which DBS either induced or alleviated psychotic symptoms may help identify circuits causally involved in psychosis and suggest candidate targets for intervention. We systematically reviewed the literature to identify all published cases in which DBS modulated (i.e., caused or improved) psychotic symptoms, regardless of target and indication. Authors of original publications were contacted to gather individual case data, allowing DBS electrode reconstruction and stimulation volume modeling. This data was aggregated into standard space and used to characterize anatomical structures most consistently associated with change in symptoms. After screening 332 studies, 36 cases were retained. This included 16 patients who received DBS for treatment-resistant schizophrenia or psychosis (nucleus accumbens, N=7; subgenual cingulate, N=4; substantia nigra pars reticulata, N=3; habenula, N=2) and 18 patients who received DBS for treatment of other conditions and experienced psychotic symptoms as a side effect (anterior nucleus of the thalamus, N=7; centromedian nucleus, N=1; subthalamic nucleus, N=6; nucleus accumbens, N=2; globus pallidus pars interna, N=1; amygdala, N=1). Finally, DBS of the nucleus basalis of Meynert improved visual hallucinations in two additional cases. Although stimulation sites were anatomically heterogeneous, qualitative integration of the empirical anatomical findings with current neurobiological models of schizophrenia revealed two circuits potentially implicated in psychotic symptoms: one centered on the mediodorsal nucleus of the thalamus and its main subcortical afferents, and one involving the nucleus accumbens - ventral tegmental area loop. We propose a preliminary theoretical framework linking these circuits to the emergence and improvement of psychotic symptoms, thereby generating testable hypotheses for future mechanistic and clinical studies. We suggest that disruption of these circuits may respectively relate to impaired filtering of cognitive and limbic representations, and aberrant salience processing.