Diagnosis of schizophrenia by integrated saccade scores and associations with psychiatric symptoms, and functioning

Abstract

Eye movement as a neurobiological biomarker of schizophrenia. We aim to estimate diagnostic accuracy of integrated pro/antisaccade eye movement measurements to discriminate between healthy individuals and schizophrenic patients. We compared the eye movement performance of 85 healthy individuals and 116 schizophrenia-stable patients during prosaccade and antisaccade tasks. The difference eye movement measurements were accumulated by stepwise discriminant analysis to produce an integrated score. Finally, the diagnostic value of the integrated score was calculated by the receiver operating characteristic (ROC) area under the curve (AUC), and the best sensitivity and specificity were calculated based on the given cutoff values. Using discriminant analysis, an integrated score included the residual gain and latency (step) during the prosaccade test, the error rate, and the corrected error rate during the antisaccade test. We found that the integrated score could well classify schizophrenia patients and healthy individuals with an accuracy of 80.6%. In the ROC, Youden's index was 0.634 (sensitivity = 81.0%, specificity = 82.4%) and AUC was 0.871. There were significant difference patterns of correlation between the severity of psychiatric symptoms and daily functioning and diagnostic eye movement measurements. Using only 2 saccade tasks to discriminate well between schizophrenia patients and healthy controls, suggesting that abnormalities in saccade behavior is a potential biomarker and efficient diagnostic tool for identifying schizophrenia. The underlying neuropathologic mechanisms associated with abnormal saccades may provide insights into the intervention and diagnosis of schizophrenia.

Figure 1.

The flow diagram of the selection.

Figure 2.

Examples of saccade tasks. Representative saccade of healthy individuals and schizophrenic patients during a single trial are shown on the left and right sides, respectively. Specifically, the overlap paradigm picture utilized in this test is shown; the central fixation (white circle) was presented at the same time as the target (white square), and the direction and sequence of saccades (white arrow) are superimposed. (A) Prosaccade test (overlap paradigm, 8°); compared with patients with schizophrenia, healthy controls required fewer eye saccades to reach the target and were more precise. (B) Antisaccade test (overlap paradigm, 12°); compared with patients with schizophrenia, healthy controls exhibited a lower error rate of antisaccades. (C) Antisaccade test (overlap paradigm, 12°): compared with patients with schizophrenia, healthy individuals exhibited a higher rate of correction following an erroneous antisaccade.

Figure 3.

The profile of eye movement scores in schizophrenia patients (SZ) and healthy controls (HC). The Y-axis represents the integrated eye movement score for each subject. HC = healthy controls, SZ = schizophrenia patients.

Figure 4.

Using ROC to predict patients with schizophrenia. AUC was 0.871, with a 95% confidence interval of 0.822 to 0.920. AUC = area under the curve, ROC = receiver operating characteristic.