These authors contributed equally to this work.
Attentional biases toward emotional scenes may represent vulnerability and maintenance factors in depression. Antidepressant therapy may improve cognitive function and reduce depression, and is considered as the mechanism of action of antidepressants. Therefore, we conducted an eye-tracking test to examine whether selective serotonin re-uptake inhibitor (SSRI) antidepressants can reduce negative attentional biases and elicit clinical responses in depression.
Twenty first-episode depressive patients freely viewed three types of pictures that depicted different emotional scenes (i.e., positive-control, neutral-control, and negative-control) for 4,000 ms while their eye movements were monitored. The attentional bias to different emotional scenes was assessed before and after eight weeks of SSRI treatment using the eye-tracking method. The control group included a group of healthy individuals.
The results revealed that first-episode depressive patients oriented their gaze more frequently to negative images and less to happy images, compared to controls. Importantly, the attentional bias in depressive patients was regulated after eight weeks of SSRI treatment. Patients showed an increased tendency to fixate on positive images and a decreased tendency to focus on negative images.
This suggests that SSRI antidepressants decrease vulnerability to negative images, while having an effect on attention in respect to positive images.
Major depressive disorder (MDD) is a serious mental disorder, characterized by persistent low mood, lowered self-evaluation, decreased motivation, insomnia, and cognitive impairment [
Recent studies using various tasks have shown that depression is characterized by an attentional bias towards negative information. Compared with healthy individuals, depressive patients showed a sustained attentional bias towards negative words that persisted throughout non-emotional processing trials [
Eye movement recordings can be considered a useful tool to assess attentional biases to visual stimuli, allowing continuous monitoring of attentional orienting and elucidating the different components of attention [
With selective serotonin reuptake inhibitors (SSRIs), a substantial proportion of patients with MDD respond to pharmacological treatment at week eight [
In accordance with previous (behavioral) findings and eye-tracking evidence in MDD, we set a number of goals. First, we examined whether attentional biases are present in depressed individuals, i.e., whether untreated depressed individuals attend less to positive images or attend more to negative images than healthy controls in terms of fixation time and/or frequency. Second, we examined whether attentional biases among depressed individuals were reversed following eight weeks of antidepressant treatment, i.e., whether less attention is paid to negative emotions, while greater attention is paid to positive emotions.
Twenty first-episode depressive patients during minor and major depressive episode were recruited from the First Affiliated Hospital of Anhui Medical University and Anhui Mental Health Center. The 20 healthy participants were recruited from local advertisements to citizens of Hefei City, Anhui Province, China. All participants were matched for age, gender, and years of education (
Gaze location and duration were measured using a remote SMI-RED eye-tracking system (SensoMotoric Instruments, Teltow, Germany), an infrared video-based eye-tracking device which allowed for recording at a frequency of 250 Hz (coordinates were sampled every 4 ms). Visual fixations were defined as eye movements lasting more than 100 ms within a 1° radius of visual angle. The SMI-RED Eye tracker was capable of compensating changes in head position, allowing the participant free head movements across a wide range, so that a head resting device was not required. Two Dell computers, one of which was served as the main test machine, were used to carry out the experimental operations, and the other desktop computer was used to present the experimental materials with a resolution of 1680×960. The tracker was placed beneath the desktop computer screen. Eye movements were monitored through the reflection of infrared light on the cornea and the pupil, which is sensed by the tracker (
The stimuli included 100 pictures (i.e., 20 positive target pictures, 20 negative target pictures, 20 neutral target pictures and 40 neutral control pictures), which were selected from the International Affective Picture System (IAPS) [
Each experiment began with a fixed cross displayed centrally on a 22-inch monitoring computer screen (screen resolution: 1920×1080 pixels), followed by a pair of two pictures, which were presented for 4,000 ms. Participants were requested to view pairs of pictures freely while recording their eye movements. After 4,000 ms of viewing the picture pairs, the pictures disappeared and the fixed cross reappeared at the same time, which indicated the commencement of the next trail. Before the formal experiment, participants were shown the task with six practical trials involving emotional and control pictures. Thus, a total of 80 trials (60 study+20 filler) were presented. Twenty filler trials with two neutral control images were presented to obscure the nature of the task. The experimental design is presented in
The study employed a case-controlled observational experiment. The sample of first-episode depressive patients (n=20) and healthy controls (n=20) were assessed at baseline using HAMD, MoCA, and an eye movement experiment. Patients were administered SSRIs each morning for a period of eight weeks without an adjustment of the dosage. In respect to the depressive patients group, assessments of attentional bias and treatment response were conducted using the eye-tracking method at baseline, as well as at week eight. Healthy controls included 20 participants who were assessed at baseline only.
In the free-viewing experiment, we computed the following eye movement indicators to reflect the participants’ attention to the pictures: 1) total fixation time (the total duration of the gaze in response to emotional pictures during the 4,000 ms free-viewing period); 2) total number of fixations (the total number of times that each participant fixed and re-fixed on a specific target picture); 3) mean glance duration (average time of each participant’s gaze was maintained within a specific boundary of a target image, i.e., the period of time between when the gaze focused on the image and when the gaze focused away from the image); 4) first-pass fixation duration (the sum of fixation duration of the image when it is first viewed) To investigate group differences in attentional bias at baseline, a 2×3 repeated-measures ANOVA was conducted, with the between-subject factor group (first-episode depression patient group at baseline, control group), and valence (positive, neutral, negative) as within-subjects factors for each eye-tracking indicator (i.e., first-pass fixation duration, total fixation number, total fixation time, and glance duration). To evaluate the therapeutic effects of SSRIs on attentional bias, first, a 2×3 repeated-measure ANOVA was conducted, with the within-subjects factors drug (baseline, eight weeks) and valence (positive, neutral, negative) in the patient group. When the interaction was significant, a paired-samples t-test was carried out to perform planned comparisons, which consisted in within-group comparisons before and after eight weeks of medication treatment. Second, we conducted a 2×3 repeated-measures ANOVA to examine the effect of SSRI treatment, with the between-subject factor group (first-episode patient group after eight weeks of treatment, control group), valence (positive, neutral, negative) as within-subjects factors for each eye-tracking indicator.
Demographic characteristics for participants according to the levels of depression, cognitive function, gender, age, and education are presented in
The ANOVA of the total number of fixations revealed the effect of valence [F (1, 38)=5.82, p=0.008] and the effect of group [F (1, 38)=6.64, p=0.014], which was qualified by an interaction of group×valence [F (2, 76)=15.40, p<0.001]. The simple effect tests on the effects of the groups for each valence revealed significant differences between the groups for the positive pictures (t=-4.73, p<0.001) and negative pictures (t=2.18, p=0.035) (
The ANOVA of the total fixation time revealed an effect of group [F (1, 38)=5.72, p=0.02]. Importantly, we found a significant group×valence interaction [F (2, 76)=27.60, p<0.001]. A simple effect analysis of the total fixation time of different emotional pictures revealed statistical difference for positive pictures between the two groups (t=-5.09, p<0.001), suggesting that depressive patients paid less attention to positive emotions compared to the control group (
The analyses showed a significant valence ×group interaction, [F (2, 76) =27.97, p<0.001]. Regarding differences between groups, simple effect tests revealed that, in comparison with the control group, patients showed a higher tendency to fixate on negative pictures (t=4.06, p<0.001) and lower tendency to fixate on positive pictures (t=-4.30, p<0.001) (
The ANOVA of the first-pass fixation duration revealed a non-significant effect for group×valence [F (2, 76)=1.88, p>0.05] (
A 2×3 repeated-measures ANOVA was conducted to assess the effects of SSRIs on the total number of fixations. A significant within-subjects interaction of drug and valence was found [F (2, 38)=8.95, p=0.003]. A paired-samples t-test revealed that more number of fixations (t=-2.46, p=0.024) for positive pictures, while less number of fixations was found for negative pictures (t=1.98, p=0.048) after eight weeks of SSRI treatment in the patient group (
The ANOVA of the effects of SSRIs in respect to changing attentional bias showed that the interaction of valence×drug was significant [F (2, 38)=14.47, p<0.001]. Following SSRI therapy (eight weeks), the fixation time for positive targets was markedly increased (i.e., paired difference between preand post-treatment: t=-3.26, p=0.004) (
The ANOVA results revealed that the interaction of valence×drug was significant [F (2, 38)=18.29, p<0.001]. To explore this difference in the patient group, a paired-samples t-test was carried out, revealing a longer fixation time for positive pictures (t=-4.33, p=0.015), while a shorter fixation time was found for negative pictures (t=2.81, p=0.01) (
The interaction of valence×drug was not significant between pre- and post-treatment in the depressive patients group [F (2, 38)=2.12, p>0.05] (
We also examined the relationship between eye-tracking parameters and clinical severity, which was measured using HAMD. Specifically, the bias scores for the negative images (i.e., the total number of fixations and glance duration) were significantly correlated with the participants’ mood symptoms (r=0.35, 0.39; p=0.039, 0.044). In addition, depressive symptoms were associated with the happy images bias in respect to the total fixation time and glance duration (r=-0.36, -0.43; p=0.03, 0.009). This is consistent with the idea that clinical severity plays a major role in the emotion-related attentional biases in depressed individuals.
The current study examined the effects of SSRIs on attentional bias in depressive patients by carrying out a free-viewing eye-tracking experiment. The key findings can be summarized as follows. First, the initial orientation of all participants (both patients and healthy controls) was determined: First-episode depressive patients oriented their gaze more frequently to negative images and less to happy images, compared to controls. Second, the attentional bias in depressive patients was regulated following eight weeks of SSRI treatment. Patients showed an increased tendency to fixate on positive pictures and a decreased tendency to fixate on negative pictures. In the SSRI-medicated depressive patients, attention to emotional pictures (positive and negative) was close to “normal” when compared to controls. Furthermore, the attentional biases of depressive patients were associated with clinical severity.
In respect to the nature of emotional biases in depressive patients, the biases toward positive and negative images were state-dependent. On the one hand, comparable to patients with bipolar disorder in the depressive phase [
Previous studies found that serotonin plays an important role in the processing of affective stimuli and inhibitory control of behavior and adaptation [
Recent imaging studies argued that depressive patients who went on to respond to SSRIs experienced a greater decrease in neutral activity, observed across a network of regions, including the anterior cingulated and amygdale [
In a previous study using a dot-probe task [
Our study illuminates the relationship between attentional bias and clinical severity in depression. Consistent with the several prior studies [
From the point of view of the relevant statistical data, after 8 weeks of treatment, total fixation time on negative image in depression showed a declining trend. However, the difference is not statistically significant. A major limitation of the current study is the relatively small sample size of 20 first-episode depressive patients, so the research results should be considered as strictly. Another limitation relates to the fact that the attentional bias in depressive patients was regulated after SSRI treatment at the eye-tracking level. Further investigations should employ a wide range of experimental tasks and designs, as well as recording methodologies in order to facilitate a more comprehensive understanding of brain networks underlying the relationship between depression and attentional processing.
In summary, our study demonstrated that first-episode depressive patients oriented their gaze more frequently to negative images and less to happy images, compared to controls. Importantly, the present study provides key evidence that following eight weeks of antidepressant treatment, the attentional bias changes in first-episode depressive patients, who showed an increasing tendency to fixate on positive pictures and a decreasing tendency to fixate on negative pictures. These observations may be relevant for the development of improved therapeutic approaches to depression.
The online-only Data Supplement is available with this article at
This research was supported by the Natural Science Foundation of China (91432301, 81771456, 81971689, 31771222 and 31571149), the Natural Science Research Project of Anhui Province (KJ2019A0239), the Humanities and Social Sciences Project of Anhui Province (SK2019A0177).
The authors have no potential conflicts of interest to disclose.
Conceptualization: Kai Wang, Chunlan Cai. Data curation: Lei Zhang, Rongrong Xuan. Formal analysis: Fengqiong Yu, Qian Hu, Yuxi Qiao. Funding acquisition: Kai Wang. Methodology: Lei Zhang, Fengqiong Yu. Project administration: Gongjun Ji, Chunyan Zhu. Resources: Chunlan Cai. Software: Kai Wang. Supervision: Kai Wang, Chunlan Cai. Validation: Lei Zhang, Qian Hu. Writing—original draft: Lei Zhang, Fengqiong Yu. Writing—review & editing: Kai Wang, Chunlan Cai.
Examples of one negative target picture and one control picture in a trial.
A: total number of fixations on each valence for pretreatment depressed patientes and control groups. B: total fixation time on each valence for pre-treatment depressed patients and control groups. C: mean glance duration on each valence for pre-treatment depressed patients and control groups. Error bars indicate standard error. *p<0.05, ***p<0.001.
A: Total number of fixations on each valence for pre- and post-treatment depressed patients. B: Total fixation time on each valence for pre- and post-treatment depressed patients. C: Mean glance duration on each valence for pre- and post-treatment depressed patients. *p<0.05, **p<0.01.
Demographic data of first-episode depression patients and controls
First-episode depression patients (N=20) | Controls (N=20) | χ2/p | |
---|---|---|---|
Gender (male/female) | 8/12 | 9/11 | 0.75 |
Age (years) | 30.5 (9.9) | 28.8 (4.4) | 0.64 |
Education (years) | 12.6 (2.2) | 14.8 (2.4) | 0.53 |
Hamilton depression scale score | 25.7 (3.5) | 5.8 (1.2) | 0.00 |
Montreal cognitive assessment score | 27.11 (0.82) | 27.67 (0.69) | 0.45 |
Mean and standard deviation for the total fixation time, the glance duration, the total number of fixations and the first-pass fixation duration for each stimulus category for depression at baseline, depression after 8 weeks treatment and control groups
Stimulus category | Depression at baseline (N=20) |
Depression after 8 weeks treatment (N=20) |
Controls (N=20) |
|||
---|---|---|---|---|---|---|
M | SD | M | SD | M | SD | |
Total fixation time (ms) | ||||||
Positive | 1,584 | 410 | 1,912 | 475 | 2,112 | 215 |
Neutral | 1,669 | 356 | 1,718 | 433 | 1,853 | 149 |
Negative | 1,839 | 423 | 1,636 | 450 | 1,731 | 167 |
Glance duration (ms) | ||||||
Positive | 1,939 | 356 | 2,329 | 379 | 2,402 | 323 |
Neutral | 2,006 | 246 | 2,086 | 295 | 1,971 | 235 |
Negative | 2,284 | 350 | 1,970 | 334 | 1,941 | 153 |
Total number of fixations | ||||||
Positive | 6.25 | 1.21 | 7.24 | 1.48 | 8.02 | 1.14 |
Neutral | 6.14 | 1.18 | 6.29 | 1.16 | 6.88 | 0.96 |
Negative | 7.23 | 1.38 | 6.50 | 1.10 | 6.82 | 0.69 |
First-pass fixation duration (ms) | ||||||
Positive | 254 | 69 | 290 | 123 | 268 | 99 |
Neutral | 234 | 74 | 276 | 170 | 281 | 145 |
Negative | 278 | 105 | 271 | 147 | 275 | 98 |
N: number, M: mean, SD: standard deviation, ms: millisecond