ERG Components in Schizophrenia and Bipolar Disorder Type I

This study will further assess ERG components obtained with different ERG devices, to be considered in a prediction model for each diagnosis. The prediction models are diaMentis proprietary software used as an ERG-based diagnostic test (classified as a Software as Medical Device, SaMD) to support the diagnosis of schizophrenia and bipolar disorder type I. They involve the processing and analysis of specific retinal biosignatures (RSPA) with the support of statistical and mathematical modelling processes e.g. machine learning and statistical learning.

The technology under development by diaMentis is defined as a Software as a Medical Device (SaMD); it will be used in combination with an electroretinogram (ERG). This study will be performed using three different ERG devices, currently marketed and cleared by the health authorities (Espion, UTAS and RETeval) to support the analytical, scientific and performance validity of the SaMD. Anomalies detected by ERG provide an objective measure that may reflect specific underlying dysfunctions in patients and thus hold promise to confirm relevant biosignatures in psychiatric disorders. Significant differences between patients with SZ, BPI and control subjects have been found despite confounding factors; this trial is required to better define the impact of patient characteristics on ERG features with a potential to refine the interpretation of results. This is a multicenter study. Three hundred subjects will be enrolled into three groups: 100 SZ patients, 100 BPI patients and 100 control subjects (healthy volunteers). The primary objective is to further characterize the ERG components in SZ and BPI patients in order to develop prediction models that discriminate each pathology. The secondary objectives are the evaluation of the repeatability and reproducibility of the analysis of the ERG components in control subjects, the assessment of the reliability of ERG prediction score for patients following a repeat test, and the evaluation of the impact of different ERG devices on the data generated and the prediction models.