Simple Eye Tests Accurately Identify Schizophrenia

Deborah Brauser

November 01, 2012

Using a series of simple eye tests to identify abnormal eye movements may help clinicians to distinguish patients with schizophrenia from those without, new research suggests.

A study of almost 90 patients with schizophrenia and almost 90 of their healthy peers showed that tests focusing on viewing patterns could detect eye-movement abnormalities associated with the disorder. And in one combined data model, the tests correctly discriminated between the 2 groups with 98.3% accuracy.

"This is a tremendously exciting result, but test sensitivity isn't enough. We want to know about specificity — whether the same test can also distinguish schizophrenia from other illnesses," lead author Philip Benson, CPsychol, AFBPsS, senior lecturer in psychology at Kings' College, University of Aberdeen, United Kingdom, told Medscape Medical News.

Dr. Philip Benson

Dr. Benson noted that his team is "very excited" about further research, which is as yet unpublished, that has actually shown that these eye tests can also distinguish patients with bipolar disorder from those with schizophrenia and from healthy control participants.

"We are quietly confident a subset of unipolar/major depression may also be accessible to our approach," he said.

For now, the investigators note that these eye tests are cheap and easy to administer in the clinical office.

"Eye-movement abnormalities appear to be stable traits, independent of current neuroleptic medication, polypharmacy, or mental state at the time of testing. They may prove to be valuable potential biological markers of the schizophrenia diathesis."

The study is published in the November 1 issue of Biological Psychiatry.

Noninvasive Biomarker

"Psychiatrists have spent a long time looking at smooth pursuit and saccadic eye movements in patients with major psychiatric illnesses," said Dr. Benson.

Although the ability of these tests to differentiate patients with schizophrenia from healthy control participants has shown promise, accuracy levels have not been impressive, with many "stuck" at the 60% to 70% level, he noted.

Dr. Benson added that both he and co–lead author David St. Clair, from the Department of Mental Health at the University of Aberdeen and from the Royal Cornhill Hospital in Aberdeen, "have a lot of experience" in research assessing eye movements and the use of electroencephalography (EEG) for event-related potential (ERP).

"It soon became clear that our different backgrounds and skills could be the ideal 'double-hander' to see if a simple, noninvasive biomarker for major psychiatric illnesses could become reality," he explained.

For this study, 88 participants with a clinical diagnosis of schizophrenia and 88 age-matched participants without the disorder were enrolled at the Ludgwig-Maximillians-University of Munich or at 1 of 3 psychiatric research sites in Scotland.

All participants underwent a series of eye tests, including smooth pursuit, free-viewing, and gaze fixation tasks.

"In smooth pursuit, people with schizophrenia have well-documented deficits in the ability to track slow-moving objects smoothly with their eyes," write the investigators in a release.

"Near Perfect" Accuracy

During a free-viewing test, patients with schizophrenia often follow an abnormal pattern with their gaze while scanning a picture. Those without the disorder tend to follow a typical scanning pattern.

A fixation task asks that an individual to keep his or her gaze steadily on an unmoving target — a request that is often difficult for those with schizophrenia to follow.

After a 9-month interval, a subgroup composed of 26 of the patients with schizophrenia and 8 of those without schizophrenia were retested. In addition, to further test predictive validity, 36 new patients with schizophrenia and 52 without the disorder were also tested.

Results showed that all of the patients with schizophrenia had abnormal performances on each eye test compared with their healthy peers.

These results stayed stable even after taking into account sex, use of medication, and whether or not the participant was a smoker.

The predictive validity on the retest and new participants was 87.8%.

"However, when we examined the whole data set of 298 assessments, a cross-validated probabilistic neural network model...could discriminate all cases from controls with near perfect accuracy at 98.3%," report the investigators.

Worth Replicating

Dr. Benson noted that differences between those with and without schizophrenia were especially clear during the tasks involving keeping eyes steady or even just viewing a series of everyday scenes.

"This was a somewhat surprising (because it suggested a generalized deficit in schizophrenia eye movements) and very important discovery because it implicated a further network of brain areas that might not be working properly in schizophrenia, in addition to pursuit and saccade pathways," he said.

However, he added that the fact that the test is not perfect is also very important. "False positives and false negatives are 'out there' — there are some patients who can perform parts of our tests quite well, and it shouldn't come as a surprise that some 'mentally well' people are misclassified because they may not have passed through the risk period for schizophrenia or other illnesses," he said.

"It also looks as if traditional statistical methods to look at the size of group differences might mask useful information about heterogeneity in these groups' eye tracking. We were quite thorough in our analyses of the data, and the measures from retest and novel volunteers reported in the article suggest the model's accuracy and predictive validity are worth trying to replicate," said Dr. Benson.

High Sensitivity

"It is encouraging to see the high sensitivity of this model for the diagnosis of schizophrenia," said John H. Krystal, MD, editor of Biological Psychiatry, who is with Yale University School of Medicine and the VA Connecticut Healthcare System in West Haven, in a release.

"It will be interesting to see the extent to which this approach enables clinical investigators to distinguish people with schizophrenia from individuals with other psychiatric disorders," he added.

According to the release, this is something Dr. Benson and Dr. St. Clair are already working on. In fact, they have already found that eye movement abnormalities "are specific to different psychiatric subgroups, another key requirement for diagnostic biomarkers."

"The next thing we want to know is when the abnormalities are first detectable and can they be used as disease markers for early intervention studies?" they note in the release.

"We are also keen to explore how best our findings can be developed for use in routine clinical practice."

Dr. Benson told Medscape Medical News that in an article published recently in Molecular Psychiatry, Thomas Insel, MD, director of the National Institute of Mental Health, and colleagues write that there is "a raft of 3,000 tests" at the disposal of general and clinical medicine.

"Yet there are almost no objective tests used in psychiatry that could externally validate diagnoses. If we can isolate further phenotypes, then our test might be a very useful instrument to assist that process," he said.

The study was supported in part by the Royal Society of London, the Millar-Mackenzie Trust, the National Institute of Mental Health Genome-Wide Association Study, and the University of Aberdeen, and by grants from the European Framework 6 and the Scottish Chief Scientist Office. The study authors have disclosed no financial relationships.

Biol Psychiatry. 2012;72:716-724. Abstract