Background: The retina is recognized as an approachable part of the brain owing to their common embryonic origin. The electroretinogram (ERG) has proved to be a valuable tool to investigate psychiatric disorders. We therefore investigated its accuracy as a tool to differentiate schizophrenia (SZ) from bipolar disorder (BP) even after balancing patients for their main antipsychotic medication.
Methods: ERG cone and rod luminance response functions were recorded in 150 patients with SZ and 151 patients with BP and compared with 200 control subjects. We created a subgroup of subjects-45 with SZ and 45 with BP-balanced for their main antipsychotic medication.
Results: A reduced cone a-wave amplitude and a prolonged b-wave latency were observed in both disorders, whereas a reduced cone b-wave amplitude was present in SZ only. Reduced mixed rod-cone a- and b-wave amplitudes were observed in both disorders. Patients with SZ were distinguishable from control subjects with 0.91 accuracy, 77% sensitivity, and 91% specificity with similar numbers for patients with BP (0.89, 76%, and 88%, respectively). Patients with SZ and patients with BP could be differentiated with an accuracy of 0.86 (whole sample) and 0.83 (subsamples of 45 patients with 80% sensitivity and 82% specificity). Antipsychotic dosages were not correlated with ERG parameters.
Conclusions: The ERG waveform parameters used in this study provided a very accurate distinction between the two disorders when using a logistic regression model. This supports the ERG as a tool that could aid the clinician in the differential diagnosis of SZ and BP in stabilized medicated patients.
Keywords: Biomarker; Bipolar disorders; Electroretinogram; Psychosis; Retina; Schizophrenia.
Copyright © 2019. Published by Elsevier Inc.