INTRODUCTION
Perinatal depression refers to depressive symptoms that emerge during infertility treatment, pregnancy, or the postpartum period, representing a prevalent and clinically relevant mental health issue. Globally, an estimated 20%-25% of women experience elevated depressive symptoms during the perinatal period, although prevalence rates vary depending on population characteristics, assessment tools, and healthcare access [1,2]. In South Korea, perinatal mood disturbances are common and frequently undertreated, particularly among women undergoing assisted reproductive treatment [3,4]. These depressive symptoms not only cause individual distress but are also associated with impairing in maternal functioning, strained partner relationships, and adverse effects on infant emotional and cognitive development [5,6]. Therefore, early identification and effective management of perinatal depression should be a public health priority.
Despite its high burden, pharmacologic treatment for perinatal depression is often underutilized due to concerns about fetal safety, teratogenic risk, and medication transfer through breast milk [7,8]. Women undergoing infertility treatment may avoid antidepressants due to concerns about treatment efficacy or reproductive outcomes [9,10], while pregnant and postpartum women frequently experience anxiety about fetal development and neonatal medication exposure during lactation [11]. Consequently, many perinatal women remain undertreated during periods of heightened vulnerability. These barriers underscore the urgent need for safe, accessible, and nonpharmacological interventions that can be flexibly applied across the perinatal continuum. Among these strategies, neuromodulation techniques, such as transcranial direct current stimulation (tDCS), have gained attention for their favorable safety profile, portability, and potential antidepressant effects [12-14].
tDCS is a non-invasive neuromodulation method that delivers low-intensity electrical currents to targeted brain regions. Clinical trials in general populations have demonstrated that tDCS can significantly reduce depressive symptoms, often within the first 1 to 2 weeks of treatment [15]. In particular, stimulation of the left dorsolateral prefrontal cortex (DLPFC), a region involved in cognitive control and emotion regulation, leads to mood improvement by modulating cortical excitability and frontolimbic network activity [16]. Open-label studies in pregnant women with depression have also reported reductions in depressive symptoms following repeated tDCS sessions without the pharmacokinetic or teratogenic risks associated with antidepressant use [13,17]. The portability and ease of use of home-based tDCS devices further support their real-world applicability, enabling flexible deployment across different stages of the perinatal period. These features suggest that tDCS may be a promising and scalable intervention to address current gaps in perinatal mental healthcare.
Despite growing evidence of the efficacy and feasibility of tDCS in alleviating depressive symptoms in perinatal populations [13,17], little is known about individual factors that may influence treatment response. Subjective health perception, defined as an individual’s overall evaluation of their health, has gained attention as a potential indicator of depressive symptom severity in various populations [18,19]. Studies suggest that negative subjective health perception is associated with heightened psychological distress, somatic symptom burden, and poor emotion regulation capacity [20,21]. Perinatal concerns about body changes, pregnancy outcomes, or maternal functioning may influence subjective health perceptions and emotional well-being [22,23]. Despite its subjective nature, perceived health status may reflect broader affective and cognitive vulnerabilities relevant to depression, providing a useful lens for understanding individual differences in treatment response.
Although the antidepressant effects of tDCS and the role of subjective health perceptions have been studied independently, their interaction in perinatal populations remains poorly understood. Given the limitations in pharmacological treatment options during pregnancy, postpartum, and infertility care, this study investigated whether baseline subjective health perception moderates the effect of home-based tDCS on depressive symptoms in perinatal women. We proposed the following hypotheses:
• Depressive symptoms would significantly decrease following a 4-week course of home-based tDCS, with effects sustained through week 8.
• Baseline subjective health perception would be associated with the overall severity of depressive symptoms.
• The trajectory of symptom changes over time would differ according to baseline subjective health perception.
METHODS
Participants
This open-label, real-world study enrolled women aged 19-50 years with mild to moderate major depressive disorder (MDD) during the perinatal period, defined as undergoing infertility treatment, currently pregnant, or within 2 months postpartum. Diagnosis was based on the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition criteria [24], with severity confirmed by standardized cut-offs: Beck Depression Inventory-II scores of 18-28 or Montgomery-Åsberg Depression Rating Scale (MADRS) scores of 14-34. This prospective observational, multicenter study included 38 women between November 2023 and March 2025 from four academic hospitals in South Korea: Seoul National University Hospital, Ilsan CHA Hospital, Bundang CHA Hospital, and Yongin Severance Hospital, including 21 who were pregnant, 9 postpartum (within 2 months), and 8 undergoing infertility treatment. Eligible participants provided written informed consent and completed validated Korean-language assessments.
Exclusion criteria included diagnoses of bipolar disorder, psychotic depression, post-traumatic stress disorder, obsessive-compulsive disorder, high risk for suicide (MADRS item 10 score ≥5, recent suicide attempt within 6 months, or per clinical judgment), seizure history, contraindications for tDCS (e.g., metallic implants or medical devices, scalp conditions preventing electrode placement), severe uncontrolled illness, recent participation in other clinical trials, or tDCS use within the past 6 months.
All study protocol was approved by the Institutional Review Board of CHA Ilsan Medical Center and conducted in accordance with the latest version of the Declaration of Helsinki and the principles of Good Clinical Practice (approval number: ICHA 2023-07-004). Written informed consent was obtained from all participants prior to their enrollment in the study.
Clinical assessments
Demographic and clinical characteristics were collected at baseline. Sociodemographic variables included age, educational level (categorized as high school graduate, college graduate, or master’s degree or higher), marital status (married, divorced, or other), and current employment status. Job-related characteristics such as monthly household income were also recorded. Obstetric information included current reproductive status (undergoing fertility treatment, currently pregnant, or postpartum), number of offspring, whether the pregnancy was planned, method of conception (natural, intrauterine insemination, or in vitro fertilization), and number of fetuses (singleton or twin). Participants also reported their subjective perception of current health status using a single question (“How would you describe your current health status?”) rated on a 5-point Likert scale (1=very healthy to 5=very unhealthy). This questionnaire was administered during the screening stage.
Depressive symptoms were assessed using the Korean version of the 20-item Center for Epidemiologic Studies Depression Scale (CES-D). This self-reported instrument evaluated the frequency of depressive symptoms over the past week using a 4-point Likert scale (0=rarely or none of the time to 3= most or all of the time). Total scores ranged from 0 to 60, with higher scores indicating more severe depressive symptoms. The CES-D was administered at four time points: baseline (V1), week 2 (V2), week 4 (V3), and week 8 (V4). All assessments were conducted by trained researchers across the four participating institutions, as illustrated in the study design flowchart (Figure 1).
tDCS intervention
Participants underwent a 4-week home-based tDCS protocol using the MINDD STIM+ (model YMS-201B+/201BS+; Ybrain Inc.), a third-grade medical device approved by the Korean Ministry of Food and Drug Safety (license no. 21-331) for home-based treatment of mild to moderate MDD. The device comprises a clinician-configured hospital docking station and portable stimulator for patient self-administration at home.
Following V1, the participants underwent a 4-week course of tDCS. Stimulation was delivered at an intensity of 2 mA for 30 minutes per session, once daily over 4 weeks, aiming for a total of 28 sessions. Participants were encouraged to complete stimulation at least five times per week, with 20 sessions as the minimum target. Saline-soaked sponge electrodes were applied via a headband, targeting the DLPFC with anodal stimulation over the left DLPFC (F3) and cathodal stimulation over the right DLPFC (F4), according to the international 10-20 electroencephalogram system. All stimulation parameters were preprogrammed at the clinic to ensure uniformity, and the device was set to the real stimulation mode. The sessions were performed independently at home, and adherence and safety were monitored using built-in sensors, remote tracking, and follow-up calls.
Statistical analysis
Descriptive statistics were used to analyze baseline characteristics. To evaluate the longitudinal effect of subjective health perception on CES-D score trajectories during the home-based tDCS treatment period, we applied Generalized Estimating Equations (GEE) with an identity link function and a normal distribution. GEE was selected for its robustness to non-normality and its capacity to handle missing data in repeated-measures designs. Independent variables included time (4 levels: V1-V4), baseline subjective health perception (5-point Likert scale, where 1=very healthy and 5=very unhealthy) and their interaction (time×subjective health perception). Participant ID was used as the subject variable, and an exchangeable working correlation matrix was specified to account for within-subject correlations across repeated measures. The reference category for time was set as V1, and the reference group for subjective health perception had the most favorable health ratings (score=1). Model fit was assessed using Quasi-likelihood under the Independence Model Criterion. Statistical significance was evaluated using Wald chi-square tests, with a p-value of <0.05 considered statistically significant. All analyses were performed using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp.).
RESULTS
Demographic, employment, and clinical characteristics of the study participants
Baseline characteristics are summarized in Table 1. A total of 38 women treated at four academic hospitals in South Korea were enrolled in this study. The mean ages of participants and their spouses was 36.7±9.4 and 36.5±4.1 years, respectively. Most participants were married (92.1%) and had completed at least a college education (78.9%). Over half (60.5%) were currently employed as were all their spouses. Reproductive status included 21 pregnant women (55.3%), 9 postpartum (23.7%), and 8 undergoing fertility treatment (21.1%). Clinical and obstetric queries indicated that 76.2% of pregnancies were planned and 61.9% were conceived naturally. One participant was carrying twins. Participants’ baseline subjective health perception varied: 10.5% rated themselves as “very healthy” (score=1), 23.7% as “somewhat healthy” (score=2), 18.4% as “neutral” (score=3), 2.6% as “somewhat unhealthy” (score=4), and 5.3% as “very unhealthy” (score=5), while 39.5% did not respond to this item. Mean CES-D scores decreased progressively over the study period: at baseline, 8.79±6.49 at baseline, 5.54±4.53 at week 2, 4.62±4.43 at week 4, and 4.47±4.26 at week 8.
Change in depressive symptoms over time
Figure 2 shows CES-D score trajectories across timepoints and health perception groups. Overall, CES-D scores decreased significantly over the 8-week period following tDCS treatment. The most marked reduction occurred between baseline (V1) and week 2 (V2), with continued improvement through week 4 (V3) and stabilization by week 8 (V4).
Influence of subjective health perception on depressive symptom severity trajectories
Table 2 presents GEE omnibus test results. The GEE model revealed a significant main effect of time (Wald χ2=90.75, degrees of freedom [df]=3, p<0.001) and subjective health perception at baseline (Wald χ2=26.41, df=4, p<0.001), as well as a significant time×health perception interaction (Wald χ2=320.18, df=9, p<0.001). These findings indicate that the trajectory of depressive symptom change was moderated by participants’ self-rated health status at baseline. Additional analyses confirmed the robustness of this effect. When reproductive stage (infertility, pregnant, postpartum) was added as a covariate, time (Wald χ2=89.97, df=3, p<0.001) and the time× subjective health perception interaction (Wald χ2=278.53, df=9, p<0.001) remained significant, whereas the time×reproductive stage interaction was not (Wald χ2=8.65, df=6, p=0.195). When subjective health perception was dichotomized into good (scores 1-2) vs. not good (scores 3-5) with reproductive stage adjustment, the time×subjective health perception interaction also remained significant (Wald χ2=17.49, df=3, p=0.001). Finally, treating nonresponders as a separate subjective health perception category yielded consistent findings (time× subjective health perception: Wald χ2=21.42, df=6, p=0.002).
As shown in Table 3 and Figure 2, participants who perceived their health as “very unhealthy” (Group 5) demonstrated the greatest reduction in CES-D scores (ΔCES-D=-10.00). Those reporting “neutral” (Group 3) health also showed substantial improvement (ΔCES-D=-8.43), while those who viewed themselves as “somewhat healthy” (Group 2) showed minimal change (ΔCES-D=-0.33). Supplementary Table 1 provides a detailed summary of the interaction effects and coefficient estimates from the GEE model.
Adherence and safety outcomes
Of the 38 participants, 21 completed at least 20 tDCS sessions over the 4-week period. Device adherence was remotely tracked using the built-in monitoring system of the MINDD STIM+unit. Adverse events were graded for severity (mild=no interference, moderate=intermittent discomfort or discontinuation, severe=requiring intervention). No serious events occurred. Tingling was most common (n=7; 18.4%), all mild. Itching (n=1; 2.6%) and transient headache (n=1; 2.6%) were mild and self-limited, while erythema in one participant (n=1; 2.6%) was moderate and led to discontinuation. Overall, home-based tDCS was well tolerated, with adverse events infrequent and mainly mild.
DISCUSSION
This study investigated the effectiveness of home-based tDCS in perinatal women with mild to moderate depression and explored the influence of baseline subjective health perception. Depressive symptoms significantly decreased over the 8-week study period, with the most pronounced reduction occurring during the first 2 weeks of active treatment. Notably, the improvements were maintained at week 8, 4 weeks after the active treatment ended. A significant interaction between time and subjective health perception indicated that individuals who viewed their health more negatively experienced a greater symptom reduction. These results suggest that subjective health perception may serve as a clinically relevant moderator of tDCS treatment outcomes, potentially reflecting differences in emotional processing, or perceived health-related agency.
In the sample, the CES-D scores declined significantly over the 8-week period, with the most marked reduction occurring during the first 2 weeks of active tDCS treatment. Additional improvement was noted through week 4, and symptom reduction was maintained at week 8 and 4 after treatment cessation, indicating both rapid onset and sustained effects. These findings suggest that tDCS is a non-pharmacological option for managing depressive symptoms in perinatal women, a population often with limited options in the use of antidepressants because of concerns related to pregnancy, lactation, or reproductive health. Biologically, tDCS is believed to exert antidepressant effects by modulating cortical excitability and neuroplasticity, particularly through anodal stimulation of the left DLPFC [25], a region involved in the top-down regulation of negative emotions and reward processing [26,27]. Mechanistic studies have suggested that tDCS influences glutamatergic and gamma-aminobutyric acid GABA-ergic neurotransmission, enhances brain-derived neurotrophic factor-dependent synaptic plasticity, and alters connectivity within the frontolimbic circuits associated with depression [28,29]. These neurophysiological effects provide a plausible basis for the antidepressant efficacy of tDCS. Clinical trials have consistently reported early onset symptom improvement with tDCS, often within the first 1 to 2 weeks of treatment [30,31]. Prior open-label studies in pregnant women with depression have demonstrated that repeated tDCS sessions could reduce depressive symptoms without the teratogenic or pharmacokinetic concerns linked to antidepressant medications [13,32]. Building on this evidence, our study demonstrated that a structured 4-week course of home-based tDCS led to rapid improvement with a durable effect lasting up to 4 weeks post-treatment. This sustained response highlights the therapeutic value of tDCS as a practical and scalable intervention in perinatal mental healthcare.
Beyond time effects, our findings demonstrated the role of subjective health perception as a potential moderator of treatment response. Participants who rated their general health poorly consistently reported higher CES-D scores, suggesting that self-perceived health status may have captured their underlying vulnerability to depression. One possible explanation is that negative perceptions of physical health may have reflected broader cognitive-affective schemas such as heightened health-related anxiety, increased somatization, or diminished interoceptive accuracy, all of which have been linked to depressive psychopathology [33,34]. Poor subjective health may have also coincided with more pronounced somatic symptoms such as sleep disturbances, low energy, or physical discomfort, which are frequently embedded in depressive symptom profiles [18,35,36]. In perinatal population, subjective health perception may also be influenced by hormonal fluctuations, altered body image, and concerns about maternal and fetal well-being, which could further intensify emotional distress [37]. Alternatively, viewing health pessimistically may have signaled a broader sense of reduced personal agency or resilience, which, in turn, contributed to feelings of helplessness and hopelessness, which are key features of depression [38]. Although subjective health ratings did not serve as objective clinical indicators, their consistent association with mood suggests that they may offer a modest value for screening and clinical risk stratification. Including a simple self-rated health item in perinatal mental health assessments may offer a pragmatic, low-burden approach for identifying individuals who could benefit from closer monitoring, even in the absence of overt medical illness.
Interestingly, participants who rated their health more negatively exhibited the greatest reduction in CES-D scores over the 8-week period. While this may have partly reflected regression to the mean or higher baseline severity, the magnitude and consistency of the effect suggest moderating role. Even after adjusting for baseline symptom levels, those with negative health perceptions showed greater treatment responsiveness. This indicated that subjective health beliefs may have captured more than symptom burden, reflecting cognitive-affective factors such as interoceptive sensitivity, somatic preoccupation, and health-related anxiety, all of which have been linked to depression and its neural substrates [39,40]. Individuals with poorer perceived health may have exhibited a stronger focus on internal bodily states processed via interoceptive networks such as the insula and anterior cingulate cortex [41,42]. Previous neuroimaging studies have shown that DLPFC-targeted tDCS alters activity in these regions [43,44], potentially restoring the top-down regulation of somatic and emotional reactivity. Therefore, individuals with higher somatic sensitivity may be more responsive to prefrontal neuromodulation. In addition, poorer perceived health may coincide with heightened stress reactivity and dysregulated hypothalamic-pituitary-adrenal axis activity [45,46], which are well-established biological pathways linking vulnerability to depression. Anodal tDCS of the DLPFC has been shown to normalize fronto-limbic connectivity and modulate stress-related circuits, potentially restoring inhibitory control over amygdala and hypothalamic responses. From a psychological perspective, individuals who perceive themselves as less healthy may also hold stronger treatment expectancy and urgency for change, leading to greater engagement and adherence, which in turn enhances neuroplastic effects of neuromodulation. Moreover, heightened emotional salience or motivation to change may enhance therapeutic engagement. Prior research suggests that those with greater perceived vulnerability responded more strongly to interventions involving bodily awareness or self-care routines [47]. Thus, subjective health perception may reflect not only baseline symptom severity but also cognitive and emotional profiles that influence neuromodulatory treatment outcomes.
This study has several limitations. First, given the open-label design without a sham control, natural remission and placebo effects cannot be excluded. The findings should therefore be interpreted as associations rather than causal effects, and randomized controlled trials are warranted to confirm efficacy. Second, modest sample size and clinical heterogeneity across perinatal stages may have introduced variability, although it also enhances real-world clinical settings. Although the main analysis relied on the original five-level SHP scale with uneven cell sizes, additional dichotomized and nonresponse-inclusive analyses confirmed that the moderation effect was consistent. This robustness supports the validity of SHP as a meaningful marker despite sample size limitations in some categories. Third, subjective health perception was measured only once using a single-item scale; repeated assessments and the inclusion of objective health indicators would strengthen future analyses. Finally, depressive symptoms were assessed through self-reporting, without clinician-rated measures or biological markers, highlighting the need for multimodal evaluation in subsequent studies.
In conclusion, this study provides preliminary evidence supporting the feasibility, safety, and sustained antidepressant effects of home-based tDCS in treating depressive symptoms in perinatal women. The observed interaction with subjective health perception suggests that individual differences in self-appraised health may meaningfully shape responsiveness to treatment. Considering the limited pharmacological options available during the perinatal period, these findings highlight that tDCS is a promising, scalable, and non-stigmatizing therapeutic alternative. Incorporating subjective health assessments into clinical screening may also help identify individuals who will benefit the most from neuromodulatory interventions. Future larger-scale controlled studies are warranted to refine targeting strategies and optimize personalized treatment approaches in this vulnerable population.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement
Print
