ABSTRACT: OBJECTIVE: Encephalopathy is a severe pathological process induced by multiple factors and diagnosed by electroencephalogram (EEG). Early diagnosis, management and treatment improves patient's prognosis. Psychotropic treatments are risk of drug-induced encephalopathies. In this study, prevalence of encephalopathies in a psychiatric hospital has been studied for 5 years (2012 to 2016) using 5217 EEG records. METHODS: EEG have been performed i) upon patient admission systematically, ii) inexplicable modifications of consciousness or behaviour or when metabolic anomalies occurred, and iii) to perform a therapeutic monitoring in external consultations. When encephalopathy was suspected, the clinical data (age, sex and concomitant treatment) and biological data (plasma levels) have been collected. Two groups were formed, namely probable encephalopathy group (with EEG signs of encephalopathy) and rejected encephalopathy group (without EEG signs of encephalopathies). RESULTS: Encephalopathy have been suspected for 189 patients and highly probable for 52 patients with a prevalence of 1% per year and rejected for 137 patients. The suspicion encephalopathy was made on both clinical (n=28) and non-clinical (n=24) signs. Involved drugs were mainly valproic acid (n=14), lithium (n=11) and clozapine (n=11) in the highly probable encephalopathy group. CONCLUSIONS: Our study demonstrates the interest of EEG in the diagnosis and monitoring of encephalopathies in a psychiatric hospital. Clinical symptoms of encephalopathies are polymorphic and sometimes atypical. Its diagnosis is underestimated in a context where behaviour or consciousness disorders are not easily imputed to psychotropic drugs used in psychiatry.