Retinal structural changes which were investigated by optical coherence tomography (OCT) have been reported in schizophrenia. Since cognitive dysfunction is a core feature of schizophrenia, the correlations between retinal findings and cognitive functions of patients and their healthy siblings may provide insight into the pathophysiological processes of the disorder. We aimed to investigate the relationship between neuropsychiatric tests and retinal changes in schizophrenia patients and their healthy siblings.
We measured OCT parameters and cognitive performance (via Trail Making Tests, verbal fluency tests, and The Digit Span Tests) of 72 participants (36 patients with schizophrenia and 36 healthy siblings) and disease severity (with Positive and Negative Syndrome Scale, Global Assessment of Functioning, and Clinical Global Impression scales) in patients with schizophrenia and evaluated the relationship between retinal findings and clinical parameters, especially neurocognitive tests.
We found decreased ganglion cell layer-inner plexiform layer thickness and macular volume in the patient group. There were strong correlations between neurocognitive tests and OCT findings in both groups. On the other hand, there was not any correlation between retinal findings and disease parameters.
The cognitive symptoms of schizophrenia may be more closely related to structural changes in the retina.
Despite a great number of research, our knowledge of the molecular pathways of schizophrenia pathophysiology remains unclear. Nevertheless, numerous structural changes such as gray matter and total brain volume decreases, reductions in the anterior cingulate cortex, corpus callosum, thalamus, and hippocampus volumes, enlargement of the lateral ventricles that were shown by neuroimaging studies suggest that biological changes at structural levels start at the earliest stages of the disease [
The retina, which is the only part of the central nervous system that allows examining with the naked eye, emerges from the forebrain ectodermal tissue during neurodevelopment, later becomes a part of the eye, and therefore it is referred to as “a window into the brain.” Optical coherence tomography (OCT) is a noninvasive, fast imaging technique that provides high-resolution and cross-sectional images of the retina. There are no contraindications to OCT, which measures specific retinal structures such as peripapillary retinal nerve fiber layer (RNFL, axon layer of retinal ganglion cells), peripapillary ganglion cell layer (GCL, bodies of ganglion cells), and inner plexiform layer (IPL, synaptic connections between the axons of bipolar cells and dendrites of ganglion cells). There is growing evidence suggesting that retinal abnormalities may be present in psychiatric disorders. Visual impairments including poor visual acuity, cataracts, strabismus, nystagmus, retinal venulae widening, various retinopathies and maculopathies, and dopamine and glutamate abnormalities in the retina are also important features of schizophrenia and could serve to define the disease state [
Additionally, there are studies that investigated cognitive performance in relation to retinal thickness. Recently, a prospective analysis of participants from the UK aged 40–69 years without neurodegenerative disease, determined that thinning of specific retinal segments was related to worse cognitive functioning, and greater likelihood of future cognitive decline. Since cognitive dysfunction is a core feature of schizophrenia, the correlation with retinal findings and cognitive functions may contribute to a better understanding of some neurodevelopmental aberrations in the disorder [
There are limited number of studies that assessed retinal layer changes in unaffected siblings of schizophrenia patients and no previous study examined the correlation of cognitive functions of the healthy siblings with retinal findings. Here, we present the first study to our knowledge, comparing the patients with schizophrenia and their unaffected siblings in terms of the relationship between retinal layer thicknesses and cognitive decline. We hypothesized that the relationship between retinal changes and cognitive parameters would be stronger in both patients and their healthy siblings, decisively.
Seventy-two participants were enrolled in the study: 36 patients fulfilling the criteria of schizophrenia according to the Diagnostic and Statistical Manual of Mental Disorders-5th edition (DSM-5) [
All respondents provided informed consent at the beginning of the survey by confirming their willingness to participate in this study. The study has been approved by the Ethics Committee of Marmara University Faculty of Medicine with the document number 2021/291 and has therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments [
Participants were questioned for age, gender, education, working situation, comorbid psychiatric or physical diseases, history of suicide attempt or ECT, body mass index (BMI), family history for psychiatric disorders, and alcohol, substance, or tobacco use. The patient group was also questioned for age at onset, duration of untreated psychosis and medications.
Positive and Negative Syndrome Scale (PANSS) was used for measuring symptom severity of patients with schizophrenia [
Global Assessment of Functioning (GAF) was used to evaluate the severity and recovery of the symptoms for the patients [
Clinical Global Impression-Severity Scale (CGI-SS) was used to score the severity of the illness and the degree of recovery. The CGI-SS is a 7-point Likert scale (1, normal; 2, borderline mentally ill; 3, mildly ill; 4, moderately ill; 5, markedly ill; 6, severely ill; and 7, among the most extremely ill patients) [
The Trail Making Test-A includes 25 circles numbered from 1 to 25 and participants are asked to draw lines to connect these numbers in ascending order. In Trail Making Test-B, the circles include both numbers and letters. Participants are asked to connect the circles using both numbers and letters in ascending order [
Categorical verbal fluency was evaluated by asking the participants to count the maximum amount of animal species in 60 seconds, while semantic fluency was evaluated by counting the words starting with the letters K, A, and S except for proper nouns in 60 seconds. The Turkish norms of the tests, the Turkish version of which was created by Öktem for the first time, were used in the study conducted by Tümaç and Canbeyli [
In the Digit Span-forward Test, patients are asked to repeat an increasing series of numbers after the tester, and the test measures short term auditory recall. In the Digit Span-backward Test which measures working memory, patients are asked to repeat the numbers in reverse order. The Turkish validity and reliability study of the scale was performed by Karakaş [
OCT is a noninvasive method that provides high-resolution images in cross-sectional imaging of the orbital tissues. By measuring the intensity of the reflected light from the tissue layers and the reflection delay time, it provides detailed tomographic section images of biological tissues at high resolution (1–15 μm) [
The Shapiro–Wilk test was used to evaluate the normality of the numeric data. Numerical variables which were normally distributed were presented as mean±standard deviation. Otherwise, presented with median (minimum–maximum) values. Categorical variables were presented with number (%) values. Independent samples t-test was used to compare normally distributed numerical variables between genders. Mann-Whitney U test was used to compare nonnormally distributed numerical variables between genders. Correlations between the numerical variables were evaluated with Spearman’s correlation coefficient or Pearson correlation coefficient according to normal distrubition. To evaluate whether retinal layers predict cognitive tests, two linear regression models were conducted. A two-sided p-value of less than 0.05 was judged to be statistically significant. Statistical analyses were performed on IBM SPSS Statistics version 23.0 (IBM Corp., Armonk, NY, USA).
Our study consisted of 36 schizophrenia patients and their 36 healthy siblings. While there were 24 men and 12 women among schizophrenia patients, this ratio was 18/18 for siblings (p=0.151). The median duration of education for patients was 8 years, while it was 11.5 years for the siblings (p=0.052). The rate of participants who are working in a full-time job was 19.4% in the patient group and 44% in the sibling group. Median BMIs were 26.16 and 22.31 kg/m2 for the patient group and their siblings, respectively (p=0.018).
For the patient group, the mean age at onset of first symptoms was 19.14±3.97 years. The median scores of the PANSS positive, negative, general psychopathology, and total were 13, 20, 33, and 66, respectively. The mean GAF and CGI-SS scores of the patients were 55 and 4, respectively. The median number of antipsychotics used by the patients was 2.0, and 27.8% of the patients were under clozapine treatment.
Sociodemographic and clinical characteristics of participants are presented in
There were significant differences between groups in right superior (p=0.026) and left inferior (p=0.021) RFNL; left superior (p=0.049), right inferior (p=0.030), and right mean (p=0.050) GCL-IPL; right (p=0.014) and left (p=0.003) MVs. There were no significant differences in other OCT parameters. The measures and comparison of the OCT parameters between schizophrenia patients and their siblings are presented in
Trail Making Test- A and B completion times of the patients were significantly longer than those of the siblings (p<0.01). The performance of the siblings was significantly better than the patients in both categorical (p<0.01) and phonemic (p<0.01) verbal fluency. While there was no difference between the groups in terms of Digit Span-forward scores, Digit Span-backward scores were significantly lower in the patient group (p<0.01).
For the patients, there were no significant correlations between OCT parameters and PANSS, CGI-SS, and GAF scores. Right and left mean RNFL was correlated with Digit Span-forward scores (p=0.001, p=0.011), which were not different from the siblings’ scores, right and left mean GCL-IPL were correlated with verbal fluency total scores (p<0.05). Other significant correlations between neurocognitive tests and OCT parameters for the groups are presented in
For the variables that were significantly related in the correlation analysis, general linear models were created by including OCT measurements, study group (patient/control), BMI, and age variables in the model to predict cognitive functions. Since GCL-IPL thicknesses differed between the two groups in our study, it was chosen as the independent variable in regression models. Additionally, verbal fluency, which was found to be significantly related to GCL-IPL thicknesses, was accepted as a dependent variable in a linear regression model. In predicting verbal fluency total scores, right and left mean GCL-IPL was still significant when group, gender, BMI, age, and interactions between group and gender were included in the model (p=0.01, p=0.03). There was also a significant group effect. The interaction effect of other variables was not significant. The effects of study group (patient/control), BMI, and age variables on verbal fluency are presented in
In our study, we hypothesized that retinal layer thickness would be decreased more in patients with schizophrenia than healthy siblings and this reduction would be related to the severity of cognitive dysfunctions. According to results, the groups were similar in terms of mean RNFL. However, all GCL-IPL thicknesses were lower in the patient group, and a mild-to-moderate correlation between phonemic verbal fluency from cognitive functions and GCL-IPL’s mean values of total and all subregion thicknesses for both eyes in patient group was demonstrated.
There are a few studies comparing patients with schizophrenia and their first-degree relatives in terms of OCT findings [
In addition to GCL-IPL changes, MVs were significantly lower in patients in our study. A study conducted with MS patients reported that MV reduction is a parameter that demostrates neuronal loss [
Deficits in neurocognitive functions are considered to be a primary symptom cluster in patients with schizophrenia [
In recent years, studies examining the relationship between retinal structures and cognitive functions have been increasing. A population-based study investigating 11,124 eyes of 6,471 subjects found that low cognitive performance was mainly associated with thinner RNFL thickness and suggested that RNFL thickness would be an indicator for cognitive performance and present the potential for prediction of future cognitive decline [
In our study, there was a strong relationship between verbal fluency performances and thickness of GCL-IPL in patient group. The verbal fluency test is used to evaluate language skills, especially speed of processing and sustained attention [
Our study has some limitations. First of all, our results may not be representative of the general population due to the relatively small size of our sample, besides; as this is a cross-sectional study, the causality is not clear. Second, we examined the relatively limited number of retinal layers in our study, so examining deeper retinal layers may help developing different perspectives on this issue. Additionally, since there was no healthy control group in the study, OCT and neurocognitive parameters could not be compared with those who were not affected by the disease or genetic vulnerability, further studies including the participants from different phases of the disorder for example patients at first psychotic episode may provide better understanding about these connections. Finally, there may be the possible effects of psychotropics used by patients and differences in BMI between groups on retinal structures. Despite these limitations, our study is the first study that evaluates OCT parameters and neurocognitive impairment in the healthy siblings of schizophrenia patients and investigates the correlation of retinal findings in both patients with schizophrenia and their healthy siblings to cognitive functions.
In conclusion, these results indicate that different structural changes in the retinal nerve are associated with some specific cognitive domains in schizophrenia patients. OCT measurements may play an essential role in tracking neurocognitive symptoms which are the important causes of disability in patients. Our findings imply that these structural changes in the retinal nerve are related to possible structural changes in the brain that cause cognitive symptoms.
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
The authors have no potential conflicts of interest to disclose.
Conceptualization: Rümeysa Taşdelen, Batuhan Ayık, Hatice Kaya. Data curation: all authors. Formal analysis: Batuhan Ayık, Hatice Kaya. Investigation: all authors. Methodology: Rümeysa Taşdelen, Batuhan Ayık, Hatice Kaya. Resources: all authors. Writing—original draft: Rümeysa Taşdelen, Batuhan Ayık, Hatice Kaya. Writing—review & editing: all authors.
None
Sociodemographic characteristics of the patients and their siblings
Variable | Schizophrenia patients (N=36) | Healthy siblings (N=36) | p |
---|---|---|---|
Age (yr) | 29.5 (25.25–37.0) | 26.0 (21.25–37.5) | 0.191 |
Gender, male | 24 (66.7) | 18 (50.0) | 0.151 |
Education time (yr) | 8.0 (5.25–11.0) | 11.5 (8.0–13.75) | 0.052 |
BMI (kg/m2) | 26.16 (22.11–29.22) | 22.31 (20.50–26.55) | 0.018 |
Smoking status, active smoker | 19 (52.8) | 13 (36.1) | 0.155 |
Working status, working | 7 (19.4) | 16 (44.4) | 0.023 |
Values are presented as median (Q1–Q3) or number (%). Two groups compared using Mann-Whitney U test for numerical variables and using Pearson chi-square test for categorical variables. BMI, body mass index
Clinical characteristics of the patients
Variable | Value (N=36) |
---|---|
Age at onset of the illness (yr) | 19.14±3.97 |
Duration of the illness (yr) | 11.77±7.38 |
Number of hospitalizations | 1.0 (1.00–2.75) |
PANSS positive symptoms | 13.0 (11.00–16.75) |
PANSS negative symptoms | 20.0 (16.25–23.75) |
PANSS general psychopathology | 33.0 (28.00–38.00) |
PANSS total | 66.0 (58.25–78.75) |
GAF | 55.0 (45.00–60.00) |
CGI-SS | 4 (4–5) |
Number of the antipsychotic drugs | 2.0 (1.00–3.75) |
Usage of clozapine treatment | 10 (27.8) |
Use of long-term antipsychotic injection | 14 (38.9) |
Values are presented as mean±standard deviation, median (Q1–Q3), or number (%). PANSS, Positive and Negative Syndrome Scale; GAF, Global Assessment of Functioning; CGI-SS, Clinical Global Impression-Severity Scale
The comparison of the RNFL, GCL-IPL, and MVs of the patients and their siblings
OCT variables | Schizophrenia patients | Healthy siblings | p |
---|---|---|---|
RNFL (μm) | |||
Right sup. |
129.14±15.39 | 136.58±12.21 | 0.026 |
Right inf. |
141.28±17.43 | 143.97±17.87 | 0.519 |
Right nasal |
102.30±19.03 | 97.39±16.07 | 0.240 |
Right temp. |
75.0 (69.25–86.75) | 74.0 (68.00–81.75) | 0.376 |
Right total |
120.50±11.72 | 123.39±10.94 | 0.283 |
Left sup. |
137.36±19.79 | 143.33±12.80 | 0.133 |
Left inf. |
140 (119–149) | 148 (134–156) | 0.021 |
Left nasal |
97.49±18.95 | 97.89±15.98 | 0.891 |
Left temp. |
76.0 (70.00–85.00) | 73.5 (71.00–79.50) | 0.538 |
Left total |
121.49±12.25 | 126.17±9.75 | 0.110 |
GCL-IPL (μm) | |||
Right sup. |
113.5 (94.25–119.75) | 116.0 (107.25–121.75) | 0.102 |
Right inf. |
114.5 (93.75–119.00) | 117.5 (11.25–124.00) | 0.03 |
Right total |
113.5 (95.00–118.75) | 117.0 (109.75–123.75) | 0.05 |
Left sup. |
112.0 (93.25–119.75) | 117.0 (108.25–124.5) | 0.049 |
Left inf. |
113.0 (94.50–121.00) | 119.5 (109.50–124.00) | 0.065 |
Left total |
112.5 (94.00–120.75) | 117.5 (109.00–124.00) | 0.052 |
MV (mm3) | |||
Right |
7.97±0.31 | 8.16±0.33 | 0.014 |
Left |
7.92 (7.71–8.11) | 8.12 (7.97–8.44) | 0.003 |
Values are presented as mean±standard deviation or median (Q1–Q3).
independent samples t-test was used;
Mann-Whitney U test was used.
RNFL, retinal nerve fiber layer; GCL-IPL, ganglion cell layer-inner plexiform layer; MV, macular volume; OCT, optical coherence tomography
Correlations between neurocognitive tests and OCT parameters in patients
VF-P | DS-F | |
---|---|---|
RNFL RM | ||
r | n.s | 0.535 |
p | n.s | 0.001 |
RNFL LM | ||
r | n.s | 0.424 |
p | n.s | 0.011 |
GCL-IPL RS | ||
r | n.s | n.s |
p | n.s | n.s |
GCL-IPL RI | ||
r | 0.418 | n.s |
p | 0.011 | n.s |
GCL-IPL RM | ||
r | 0.371 | n.s |
p | 0.026 | n.s |
GCL-IPL LS | ||
r | 0.370 | n.s |
p | 0.026 | n.s |
GCL-IPL LI | ||
r | 0.365 | n.s |
p | 0.029 | n.s |
GCL-IPL LM | ||
r | 0.374 | n.s |
p | 0.025 | n.s |
VF-P, verbal fluency phonemic; DS-F, Digit Span forward; RNFL, retinal nerve fiber layer; RNFL RM, RNFL right mean; RNFL LM, RNFL left mean; GCL-IPL, ganglion cell layer-inner plexiform layer; GCL-IPL RS, GCL-IPL right superior; GCL-IPL RI, GCLIPL right inferior; GCL-IPL RM, GCL-IPL right mean; GCL-IPL LS, GCL-IPL left superior; GCL-IPL LI, GCL-IPL left inferior; GCL-IPL LM, GCL-IPL left mean; n.s, not significant
The effects of study group (patient/control), BMI, and age variables on verbal fluency scores
F | p | Partial | |
---|---|---|---|
Model-1 | 4.434 | 0.001 | 0.290 |
GCL-IPL RM | 5.834 | 0.019 | 0.082 |
Group | 7.609 | 0.008 | 0.105 |
Gender | 2.335 | 0.131 | 0.035 |
BMI (kg/m2) | 0.523 | 0.472 | 0.008 |
Age (yr) | 2.088 | 0.153 | 0.031 |
Group×gender | 1.498 | 0.225 | 0.023 |
Model-2 | 4.133 | 0.001 | 0.276 |
GCL-IPL LM | 4.441 | 0.039 | 0.064 |
Group | 7.668 | 0.007 | 0.106 |
Gender | 2.244 | 0.139 | 0.033 |
BMI (kg/m2) | 0.477 | 0.492 | 0.007 |
Age (yr) | 2.370 | 0.129 | 0.035 |
Group×gender | 1.321 | 0.255 | 0.020 |
Model-1, Linear regression model-1; Model-2, Linear regression model-2. GCL-IPL RM, ganglion cell layer-inner plexiform layer right mean; GCL-IPL LM, ganglion cell layer-inner plexiform layer left mean; BMI, body mass index