| Objective To acquire data of eye movements in first-episode patients with schizophrenia,first-episode patients with depression and healthy participants and compare the characteristic pattern of eye movements between schizophrenia and depression during some simple eye movement tasks using eye movement apparatus.Also,to investigate the predicative validity of eye movement parameters discriminating schizophrenia populations from depression population and healthy controls using support vector machine,a machine learning.Methods All enrolled participants were subjected to the M.I.N.I.(the MINI-International Neuropsychiatric Interview)interview and completed eye movement test.Four eye movement paradigms of smooth pursuit,fixation stability,pro/anti saccadic,free-view tasks were included.A total of 316 subjects comprising 102 schizophrenia,113 depression and 101 healthy subjects were included for final analysis.Analysis of covariance with age,education degree as covariates was applied to examine characteristics of eye movement across three groups in separate four paradigms.PCA-SVM(Principal component analysis and support vector machine)method was conducted to modeling differential diagnosis so as to differentiate schizophrenia patients from both normal and depression population via eye movement indices.Results During smooth pursuit,for horizontal task,a trend significant difference on horizontal gain was shown in schizophrenia,but not in depression relative to healthy controls,number of saccade for schizophrenia was significantly more frequent than healthy controls and depression patients,while no difference was shown between the latter two populations,no significant difference of position difference was displayed between three groups;For Lissajous curve(Hz0.2)task appeared worse performance on level gain,number of saccade and position difference comparing to healthy participant and depressio;For(Hz0.4)task,both schizophrenia and depression appeared worse performance on level gain,number of saccade and position difference comparing to healthy participant,while no significant difference between patients groups;for vertical task,there was no any significant difference on all eye movement parameters between three groups.During saccade and antisaccade paradigm,amplitude in schizophrenia was significant greater than healthy controls in spite of interference.On the contrast,only with intervention but not without intervention,depression did exhibit abnormal saccade amplitude.With intervention,the greatest saccade amplitude existed in schizophrenia,the smallest was in healthy subjects and the middle was in depression;antisaccade latency of schizophrenia was significantly longer than depression,and antisaccade latency of depression was statistically longer than healthy subjects.During free-view paradigm,schizophrenia showed abnormal eye movement pattern.Longer average fixation time,decreased number of saccadic,smaller saccade amplitude and shorter span length was observed in schizophrenia relative to healthy participants;depression did not present abnormal saccade amplitude and span length although longer average fixation time and fewer number of saccade.PCA-SVM was used for modeling of differential diagnosis.Under the condition of separating schizophrenia from depression and normal subjects,the predictive accuracy,susceptibility and specificity was 82.5%,80% and 75% respectively.if the condition was to distinguish schizophrenia from normal subjects,the predictive accuracy,susceptibility and specificity increased to be 87.5%,90% and 90% respectively。Conclusions Abnormal eye movements pattern presented in either schizophrenia or depression,as well as more obvious in schizophrenia comparing to depression in three tasks of smooth pursuit,fixation stability,anti saccadic.Schizophrenia characterized limited free-view pattern,slowed processing information and narrowing visual searching range,however,only slowed procession of information rather than limitation pattern was demonstrated in depression.Eye movements could be used for diagnosis and differential diagnosis of schizophrenia with higher accuracy and studies on eye movements are of great value. |