Longitudinal changes in brain structure and their relationship with subclinical psychiatric symptoms in parents who lost their only child in China

ABSTRACT

Background:

Losing an only child (Shidu) is a grievous traumatic event that may affect brain structure, even if it does not lead to psychiatric disorders. However, longitudinal changes in brain structure and their relationship to subclinical psychiatric symptoms (SPS) have not been well investigated in Shidu parents without any psychiatric disorders (SDNP).

Objectives:

This study aimed to investigate cross-sectional and longitudinal changes in cortical thickness and surface area in SDNP, and to explore their relationship with SPS.

Methods:

A total of 50 SDNP and 40 matched healthy controls (HC) were enrolled. All participants underwent structural MRI scans and clinical assessment at baseline and at the 5-year follow-up. Differences in brain structural phenotypes (cortical thickness, surface area, and their annual rate of change) between the SDNP and HC groups were compared using FreeSurfer. Correlations between significant brain structural phenotypes and SPS in the SDNP group were evaluated using multiple linear regressions.

Results:

The SDNP group showed a smaller surface area in the left inferior parietal cortex than the HC group at baseline and follow-up. The SDNP group showed slower rates of cortical thinning and surface area loss in several brain regions than the HC group from baseline to follow-up. Moreover, slower rates of cortical thinning in the left insula, superior frontal cortex, and superior temporal cortex were associated with greater reductions in avoidance, depression, and trauma re-experiencing symptoms scores over time in the SDNP group, respectively.

Conclusions:

Shidu trauma-induced structural abnormalities in the inferior parietal cortex may persist over time and be independent of the severity of psychiatric symptoms. The expansion of prefrontal, temporal, and insular cortex implicated in emotional regulation may contribute to improvements in psychiatric symptoms in Shidu parents.

HIGHLIGHTS

• This study focused on longitudinal changes in cortical thickness and surface area and their relationship with subclinical psychiatric symptoms in Shidu parents without any psychiatric disorders.

• Shidu trauma-induced structural abnormalities in the inferior parietal cortex may persist over time and be independent of the severity of psychiatric symptoms.

• The expansion of prefrontal, temporal, and insular cortex implicated in emotional regulation may contribute to improvements in psychiatric symptoms in Shidu parents.

Antecedentes: La pérdida de un hijo único (Shidu) es un evento traumático grave que puede afectar la estructura del cerebro, incluso si no conduce a trastornos psiquiátricos. Sin embargo, los cambios longitudinales en la estructura del cerebro y su relación con los síntomas psiquiátricos subclínicos (SPS) no se han investigado bien en los padres de Shidu sin ningún trastorno psiquiátrico (SDNP).

Objetivos: Este estudio tuvo como objetivo investigar los cambios transversales y longitudinales en el grosor cortical y el área de superficie en SDNP, y explorar su relación con SPS.

Métodos: Se inscribieron un total de 50 SDNP y 40 controles sanos (CS) emparejados. Todos los participantes se sometieron a resonancias magnéticas estructurales y evaluación clínica al inicio y en el seguimiento de 5 años. Las diferencias en los fenotipos estructurales del cerebro (grosor cortical, área de superficie y su tasa de cambio anual) entre los grupos SDNP y CS se compararon utilizando FreeSurfer. Las correlaciones entre fenotipos estructurales cerebrales significativos y SPS en el grupo SDNP se evaluaron mediante regresiones lineales múltiples.

Resultados: El grupo SDNP mostró un área de superficie más pequeña en la corteza parietal inferior izquierda que el grupo CS al inicio y en el seguimiento. El grupo SDNP mostró tasas más lentas de adelgazamiento cortical y pérdida de superficie en varias regiones del cerebro que el grupo CS desde el inicio hasta el seguimiento. Además, las tasas más lentas de adelgazamiento cortical en la ínsula izquierda, la corteza frontal superior y la corteza temporal superior se asociaron con mayores reducciones en las puntuaciones de síntomas de evitación, depresión y síntomas de reexperimentación del trauma a lo largo del tiempo en el grupo SDNP, respectivamente.

Conclusiones: Las anomalías estructurales inducidas por el trauma de Shidu en la corteza parietal inferior pueden persistir en el tiempo y ser independientes de la gravedad de los síntomas psiquiátricos. La expansión de la corteza prefrontal, temporal e insular implicada en la regulación emocional puede contribuir a mejorar los síntomas psiquiátricos en los padres Shidu.

背景：失去独生子女 (失独) 是一种严重的创伤事件，可能会影响大脑结构，即使它不会导致精神障碍。 然而，大脑结构的纵向改变及其与亚临床精神症状（SPS）的关系尚未在没有任何精神障碍的失独父母（SDNP）中充分研究。

目的：本研究旨在考查SDNP 皮质厚度和表面积的横断面和纵向改变，并探讨它们与 SPS 的关系。

方法：共纳入 50 名 SDNP 和 40 名匹配的健康对照（HC）。所有参与者在基线和 5 年随访时都接受了结构 MRI 扫描和临床评估。使用 FreeSurfer 比较 SDNP 和 HC 组之间大脑结构表型（皮质厚度、表面积及其年变化率）的差异。 使用多元线性回归评估 SDNP 组中重要脑结构表型与 SPS 之间的相关性。

结果：在基线和随访时，SDNP 组的左下顶叶皮层表面积小于 HC 组。从基线到随访，SDNP 组在几个脑区的皮质变薄和表面积损失率比 HC 组慢。此外，在 SDNP 组中，随着时间的推移，左侧脑岛、上额叶皮层和上颞叶皮层的皮质变薄速度较慢，分别与回避、抑郁和创伤再体验症状评分降低幅度更大有关。

结论：失独创伤引起的下顶叶皮层结构异常可能会随时间推移持续存在，并且与精神症状的严重程度无关。与情绪调节有关的前额叶、颞叶和岛叶皮层的扩张可能有助于改善失独父母的精神症状。

KEYWORDS:

• Trauma stress
• Shidu
• lose only child
• subclinical psychiatric symptoms
• longitudinal study
• cortical thickness and surface area

PALABRAS CLAVE:

• Estrés traumático
• Shidu
• pérdida de hijo único
• síntomas psiquiátricos subclínicos
• estudio longitudinal
• espesor cortical y área de superficie

关键词:

• 创伤应激
• 失独
• 失去独生子女
• 亚临床精神症状
• 纵向研究
• 皮质厚度和表面积

1. Introduction

To control rapid population growth, China introduced the family planning policy in the early 1970s, which encouraged a couple to have only one child; there were over one million one-child families as a result of the policy (Hesketh et al., 2005). When the only child died and their parents were unable to have another child because of age or other reasons, the family was regarded as a ‘Shidu family’ (Song, 2014) (Shidu means ‘losing the only child’ in Chinese). Shidu is an unquestionably grievous traumatic event in traditional Chinese culture, symbolizing the inability to carry on the family line and the loss of old-age security (Wang & Hu, 2021). Shidu parents close themselves off and avoid social interaction for a long time, leading to increasingly serious psychological problems and predisposition to various degrees of psychiatric disorders (Yin et al., 2018).

Previous neuroimaging studies on Shidu trauma have focused mainly on parents with psychiatric disorders (Luo et al., 2019; Qi et al., 2020). However, not every Shidu parent is diagnosed with psychiatric disorders (Yin et al., 2018). Many people do not meet the diagnostic criteria for psychiatric disorders, but they may still experience symptoms of subclinical anxiety, depression, and post-traumatic stress disorder (PTSD), which may also affect social function and work quality and increase the risk of developing psychiatric disorders (Yin et al., 2018; Zhou et al., 2018; Eli et al., 2021). A previous study by our team also discovered a decrease in hippocampal volume in this population (Luo et al., 2016). Therefore, research on Shidu parents without any psychiatric disorders (SDNP) is crucial.

Cortical thickness and surface area, which are two other measures of structural brain changes, can provide more specificity of brain-related changes than gray matter volume and have been widely used to reveal the impact of trauma on the nervous system (Busso et al., 2017). Previous studies have found associations between trauma and these two measures; however, the results have been inconsistent. Some studies reported that traumatized individuals had higher cortical thickness in the prefrontal (Jeong et al., 2021) or temporal (Nilsen et al., 2016) regions than healthy controls (HC), while another reported widespread cortical thickness or surface area reductions (Kelly et al., 2013). One reason for these inconsistencies could be that multiple trauma factors were included in these studies, and different trauma types may have different effects on brain structure (Li et al., 2022). Thus, previous research findings may not apply to Shidu, which is a severe trauma with Chinese characteristics.

Subclinical psychiatric symptoms (SPS) in traumatized individuals may improve or worsen over time without treatment, and longitudinal studies can further investigate whether the clinical trend is related to dynamic changes in the brain structure. Previous research found that earthquake survivors experienced an improvement in SPS as well as an increase in the integrity of white matter microstructure during a 2-year follow-up after trauma (Meng et al., 2018); however, the study only focused on anxiety and depression symptoms, ignoring PTSD symptoms, which often coexist in traumatized individuals. Additionally, significant brain structural changes detected by MRI need adequate time since trauma, and the average follow-up interval of two years may be relatively short, which could not lead to more distinctive alterations. Another study with a longer follow-up period (five years) found that the increased prefrontal cortex of subway disaster survivors gradually normalized over time; the greater the normalization, the greater the remission of PTSD symptoms (Lyoo et al., 2011). However, the study only included patients with PTSD, not subclinical individuals. Moreover, different scanning intervals between HC participants and traumatized individuals may also undermine the confidence of longitudinal comparisons, while the annual rate of change (Abé et al., 2022) (i.e. the annual percentage change relative to the baseline brain structure measurement) can effectively rule out the influence of scanning intervals and healthy aging to better account for dynamic changes in brain structure after trauma.

Therefore, we used a 5-year naturalistic longitudinal study to compare group differences in brain structural phenotypes (cortical thickness, surface area, and their annual rate of change) between the SDNP and HC groups, and to further evaluate correlations between significant brain structural phenotypes and SPS in the SDNP group. We hypothesized that the SDNP group would show abnormalities in emotion-processing regions at baseline, that some of these abnormalities would recover over time, and that brain changes would be correlated with improvements in SPS.

2. Material and methods

2.1. Participants

From September 2016 to March 2017, 61 SDNP and 46 sex-, age-, and education-matched HC participants were collected from Jiangsu Province, China, and followed up approximately five years later in the natural state of no clinical intervention. All participants underwent structural MRI scans and clinical scale assessment at baseline and follow-up.

Participants in the SDNP group reported no significant traumatic events other than losing their only child, and participants in the HC group reported no significant history of traumatic events. The exclusion criteria for all groups were as follows: diagnosis of psychiatric disorders (including PTSD, anxiety, depression, etc.), left-handedness, alcohol or substance dependency, head injury with loss of consciousness, neurological disorders, and MRI contraindications.

Participants were excluded if they had sustained psychological trauma, craniocerebral injury, use of antipsychotic drugs, severe physical and mental illness during follow-up, or if they were lost to follow-up. A total of 11 SDNP and 6 HC participants were excluded at follow-up. Finally, a total of 50 SDNP (23 males and 27 females; aged 47 to 65 years at baseline; mean [SD] age, 55.56 [4.32] years) and 40 sex-, age-, and education-matched HC participants (14 males and 26 females; aged 45 to 65 years at baseline; mean [SD] age, 56.20 [5.35] years) were included in the study. None of the SDNP group was diagnosed with psychiatric disorders at follow-up.

This study was approved by the Medical Ethics Committee of the Affiliated Yixing Hospital of Jiangsu University and all participants provided written informed consent.

2.2. Clinical assessment

The Life Stress Events Checklist (revised version) (Wolfe et al., 1996) was used to assess significant history of traumatic events for all participants. The Clinician-Administered PTSD Scale (CAPS) for DSM-IV (Blake et al., 1995) was used to assess the current diagnosis of PTSD and symptom severity, including the CAPS total score and its three clusters of symptoms B (trauma re-experiencing), C (avoidance), and D (hyperarousal) scores. The Hamilton Anxiety Rating Scale (HAMA) (Hamilton, 1959) and Hamilton Depression Rating Scale (HAMD) (Hamilton, 1960) were used to assess the symptoms of anxiety and depression. The higher the score, the more severe the symptom.

2.3. Image acquisition

MRI examinations were performed at 3.0 T using an eight-channel phased-array coil (Achieva 3.0 T TX; Philips, Amsterdam, the Netherlands). All participants were placed in the supine position, their eyes were closed, and their heads were fixed with foam pads to minimize motion artifacts. Three-dimensional high-resolution T1-weighted structural brain images were acquired in the sagittal orientation using the turbo fast echo, and the parameters were set as follows: repetition time = 9.7 ms, echo time = 4.6 ms, inversion time = 900 ms, flip angle = 9°, field of view = 256 mm × 256 mm, matrix size = 256 × 256, voxel size = 1 mm × 1 mm × 1 mm, number of slices = 160. The total scan time was 5 min 34 s.

2.4. Image processing

Cortical reconstruction and volumetric segmentation at each imaging time point were performed using the FreeSurfer (version 6.0; http://surfer.nmr.mgh.harvard.edu/) image analysis suite. The images were first processed using standard streams, including motion correction, removal of non-brain tissue, segmentation of white and gray matter, intensity normalization, tessellation of the gray/white matter boundary, automated topology correction, and surface deformation. After running all participants through the standard processing stream, the data were manually inspected and edited for gray/white and gray/pial surface accuracies.

Subsequently, the images were processed using a longitudinal stream (Reuter et al., 2012). Specifically, an unbiased within-subject template space and image was created based on the two cross-sectional images created for each participant using robust, inverse-consistent registration. Several processing steps, such as skull stripping, Talairach transforms, spherical surface maps, and atlas-based parcellations, were initialized with common information from the within-subject template, significantly increasing the reliability and statistical power.

2.5. Statistical analysis

Demographic and clinical variables were compared using SPSS 25.0. Independent-sample t-tests or χ² tests were used to assess between-group differences. Paired-sample t-tests were used for intragroup differences. The threshold for statistical significance was set at p < .05.

Differences in brain structural phenotypes between the SDNP and HC groups were compared using a whole-brain vertex-wise general linear model approach performed by Qdec in Freesurfer software, with sex and age as covariates. Given the effects of aging on brain structure (Sele et al., 2021) and the fact that the MRI scanning interval of all participants was not completely consistent, to minimize these potential effects on results, longitudinal changes in cortical thickness and surface area for each hemisphere were calculated as an annualized measure of the percent change [i.e. annual rate of change = (measured at follow-up − measured at baseline) × 100% / (measured at baseline × scanning interval)], where negative values reflected a decrease (i.e. cortical thinning or surface area loss) and positive values an increase (i.e. cortical thickening or surface area expansion) over time (Abé et al., 2022). Monte Carlo corrected for multiple comparisons, vertex-wise cluster threshold of 1.3, p < .05 (Hagler et al., 2006). When significant clusters were found, they were mapped to the Desikan-Killiany atlas (Desikan et al., 2006) and the corresponding brain structural phenotype measurements were extracted and exported to SPSS 25.0.

Correlations between significant brain structural phenotypes and SPS in the SDNP group were evaluated using multiple linear regression models. Each significant cluster was separately modeled as a dependent variable. Cross-sectional cortical thickness and surface area models included covariates of non-interest (sex and age) and the main predictors of interest (i.e. SPS scores, including HAMA, HAMD, B, C, and D). Longitudinal annual rate of change models included covariates of non-interest (sex and age at baseline) and the main predictors of interest (i.e. longitudinal changes in SPS scores, including ΔHAMA, ΔHAMD, ΔB, ΔC, and ΔD), where Δ = (measured at baseline – measured at follow-up) / scanning interval. To avoid an inflated Type 1 error rate, the Bonferroni method was used to correct for multiple comparisons, and the corrected threshold was p < .05/n (n reflects the number of regressions, here is also the number of brain regions with significant differences). The corrected threshold of cross-sectional regression models was p < .05, the corrected threshold of longitudinal regression models was p < .006.

3. Results

3.1. Demographic data and clinical comparisons

The demographic and clinical data of the participants are summarized in Tables 1 and 2. At both baseline and follow-up, the SDNP group had significantly higher HAMA and HAMD scores than the HC group. Neither group differed significantly in terms of sex, age, or education. However, the SDNP group had significantly shorter scanning interval than the HC group (Table 1).

Table 1. Demographic data and clinical comparisons between groups.

	SDNP (n = 50)	HC (n = 40)	χ^2/t	p
Baseline
Female (%)	27(54%)	26(65%)	1.111	.292
Age (years)	55.56 ± 4.32	56.20 ± 5.35	−0.628	.531
Years of education	7.74 ± 3.38	7.45 ± 3.67	0.389	.698
Time duration since trauma (months)	97.96 ± 59.22	−	−	−
HAMA	4.72 ± 3.32	2.00 ± 2.08	4.748	<.001
HAMD	5.84 ± 3.48	1.85 ± 1.46	7.344	<.001
Follow-up
Scanning interval (years)	4.90 ± 0.49	5.28 ± 0.08	−5.352	<.001
HAMA	5.10 ± 2.87	2.58 ± 1.62	5.270	<.001
HAMD	5.48 ± 3.25	2.05 ± 1.30	6.819	<.001

Note: Values are expressed as the mean ± SD. SDNP: Shidu parents without any psychiatric disorders; HC: healthy controls; HAMA: Hamilton Anxiety Scale; HAMD: Hamilton Depression Scale; –: not applicable.

Table 2. Clinical data comparisons between baseline and follow-up within groups.

	SDNP (n = 50)		t	p	HC (n = 40)		t	p
	Baseline	Follow-up			Baseline	Follow-up
HAMA	4.72 ± 3.32	5.10 ± 2.87	−0.762	.450	2.00 ± 2.08	2.58 ± 1.62	−1.511	.139
HAMD	5.84 ± 3.48	5.48 ± 3.25	0.752	.456	1.85 ± 1.46	2.05 ± 1.30	−1.433	.160
CAPS	17.60 ± 10.43	15.76 ± 7.21	1.176	.245	−	−	−	−
B	7.94 ± 5.01	5.16 ± 2.75	3.758	<.001	−	−	−	−
C	5.16 ± 5.13	4.00 ± 3.18	1.338	.187	−	−	−	−
D	4.50 ± 4.35	6.60 ± 4.35	−3.303	.002	−	−	−	−

Note: Values are expressed as the mean ± SD. SDNP: Shidu parents without any psychiatric disorders; HC: healthy controls; HAMA: Hamilton Anxiety Scale; HAMD: Hamilton Depression Scale; CAPS: Clinician-Administered PTSD Scale; B: trauma re-experiencing; C: avoidance; D: hyperarousal; –: not applicable.

From baseline to follow-up, the B score significantly decreased and the D score significantly increased in the SDNP group, but the C score and CAPS total score did not change significantly. The HAMA and HAMD scores in both groups did not change significantly (Table 2).

3.2. Whole-brain vertex-wise analyses

3.2.1. Group differences in cortical thickness and surface area

At both baseline and follow-up, the SDNP group had a significantly smaller surface area than the HC group, predominantly in the left inferior parietal cortex (Table 3, Figure 1). No significant group differences were found in cortical thickness.

Figure 1. Distribution maps of group differences in surface area. (A): The blue regions represent smaller surface area of the left hemisphere in the SDNP group than in the HC group at baseline. (B): The blue regions represent smaller surface area of the left hemisphere in the SDNP group than in the HC group at follow-up. The color scale is represented as −log P. SDNP: Shidu parents without any psychiatric disorders; HC: healthy controls.

Table 3. Group differences in surface area.

SDNP vs. HC	Cluster	Peak z-value	Cluster size (mm^2)	p value, corrected	MNI coordinates (x, y, z)	Region of peak z-value
Baseline	1	−3.807	1107.52	.003	−36, – 82, 27	Left inferior parietal cortex
Follow-up	1	−3.570	851.90	.014	−37, – 82, 26	Left inferior parietal cortex

Note: SDNP: Shidu parents without any psychiatric disorders; HC: healthy controls; MNI: Montreal Neurological Institute.

3.2.2. Group differences in the annual rate of change

From baseline to follow-up, the SDNP group showed a significantly slower rate of surface area loss than the HC group, predominantly in the right superior parietal cortex (Tables 4 and 5, Figure 2); the SDNP group showed a significantly slower rate of cortical thinning than the HC group, predominantly in the left superior frontal cortex, insula, pars triangularis, superior temporal cortex, and right caudal middle frontal cortex, insula (Tables 4 and 5, Figure 3).

Figure 2. Distribution maps of group differences in the annual rate of change in surface area. The yellow regions represent slower rates of surface area loss of the right hemisphere over time in the SDNP group than in the HC group. The color scale is represented as −log P. Detailed statistical results and numerical values are provided in Tables 4 and 5. SDNP: Shidu parents without any psychiatric disorders; HC: healthy controls.

Figure 3. Distribution maps of group differences in the annual rate of change in cortical thickness. (A): The red and yellow regions represent slower rates of cortical thinning of the left hemisphere over time in the SDNP group than in the HC group; (B): The red and yellow regions represent slower rates of cortical thinning of the right hemisphere over time in the SDNP group than in the HC group. The color scale is represented as −log P. Detailed statistical results and numerical values are provided in Tables 4 and 5. SDNP: Shidu parents without any psychiatric disorders; HC: healthy controls.

Table 4. Group differences in the annual rate of change.

SDNP vs. HC	Annual rate of change	Cluster	Peak z-value	Cluster size (mm^2/ mm)	p value, corrected	MNI coordinates (x, y, z)	Region of peak z-value
Baseline to Follow-up	Surface area	1	2.864	412.89	.035	16,−84, 37	Right superior parietal cortex
	Cortical thickness	1	3.886	814.97	.001	−7,−11, 55	Left superior frontal cortex
		2	2.869	596.75	.023	−35, 13,−4	Left insula
		3	2.532	581.55	.027	−52, 28, 7	Left pars triangularis
		4	3.325	570.26	.031	−20,−5, 53	Left superior frontal cortex
		5	3.898	529.50	.046	−63,−17, 1	Left superior temporal cortex
		6	4.067	817.69	.007	25, 1, 46	Right caudal middle frontal cortex
		7	3.341	664.16	.028	37, −13,−9	Right insula

Note: SDNP: Shidu parents without any psychiatric disorders; HC: healthy controls; MNI: Montreal Neurological Institute.

Table 5. Corresponding annual rate of change (%) in regions with significant group differences.

Annual rate of change	Cluster	SDNP (n = 50)	HC (n = 40)
Surface area	1	−0.2275 ± 0.3936	−0.2745 ± 0.3716
Cortical thickness	1	−0.0881 ± 0.7628	−0.2086 ± 0.4751
	2	−0.2205 ± 0.6412	−0.2798 ± 0.4464
	3	−0.3661 ± 0.6374	−0.4425 ± 0.5994
	4	−0.0245 ± 0.7011	−0.3634 ± 0.6831
	5	−0.3164 ± 0.6924	−0.3402 ± 0.5590
	6	−0.1841 ± 0.7775	−0.3157 ± 0.5306
	7	−0.1827 ± 0.8403	−0.3137 ± 0.6646

Note: Values are expressed as the mean ± SD. SDNP: Shidu parents without any psychiatric disorders; HC: healthy controls.

3.3. Regression analysis

3.3.1. Correlations between cortical thickness or surface area and SPS scores

At baseline and follow-up, the surface area in the left inferior parietal cortex was not significantly associated with SPS scores in the SDNP group.

3.3.2. Correlations between the annual rate of change and longitudinal changes in SPS scores

A slower rate of cortical thinning in the left insula was significantly associated with a greater reduction in C score (r = 0.431, p = 0.002), a slower rate of cortical thinning in the left superior frontal cortex was significantly associated with a greater reduction in HAMD score (r = 0.637, p < 0.001), and a slower rate of cortical thinning in the left superior temporal cortex was significantly associated with a greater reduction in B score (r = 0.468, p = .005) over time in the SDNP group (Table 6, Figure 4). No significant correlations were found between rates of change in cortical thickness or surface area in other regions and longitudinal changes in SPS scores.

Figure 4. Scatter plot of correlation between the annual rate of change in cortical thickness and changes in SPS scores in the SDNP group. (A): A slower rate of cortical thinning in the left insula was associated with a greater reduction in C score; (B): A slower rate of cortical thinning in the left superior frontal cortex was associated with a greater reduction in HAMD score; (C): A slower rate of cortical thinning in the left superior temporal cortex was associated with a greater reduction in B score. SDNP: Shidu parents without any psychiatric disorders; HAMD: Hamilton Depression Scale; B: trauma re-experiencing; C: avoidance; Δ = (measured at baseline – measured at follow-up) / scanning interval.

Table 6. Multiple linear regression for longitudinal changes in SPS scores and the annual rate of change in cortical thickness in the SDNP group.

Predictors	β	t	p	Adjusted R^2
Left insula (Clusters 2)				0.283
sex	0.027	0.211	.834
Age	−0.466	−3.623	.001
ΔHAMA	−0.090	−0.502	.619
ΔHAMD	0.004	0.025	.980
ΔB	−0.030	−0.205	.838
ΔC	0.431	3.306	.002
ΔD	−0.020	−0.148	.883
Left superior frontal cortex (Clusters 4)				0.303
sex	0.044	0.350	.728
Age	−0.263	−2.074	.044
ΔHAMA	−0.334	−1.880	.067
ΔHAMD	0.637	3.800	<.001
ΔB	−0.054	−0.373	.711
ΔC	−0.239	−1.862	.070
ΔD	−0.061	−0.455	.652
Left superior temporal cortex (Clusters 5)				0.171
sex	−0.039	−0.285	.777
Age	−0.162	−1.170	.249
ΔHAMA	0.186	0.960	.342
ΔHAMD	−0.104	−0.569	.572
ΔB	0.468	−2.950	.005
ΔC	−0.220	−1.570	.124
ΔD	−0.072	−0.489	.627

Note: Only the results that survived multiple comparison corrections using the Bonferroni method (p < .006) are presented. SPS: subclinical psychiatric symptoms; SDNP: Shidu parents without any psychiatric disorders; HAMA: Hamilton Anxiety Scale; HAMD: Hamilton Depression Scale; B: trauma re-experiencing; C: avoidance; D: hyperarousal; Δ = (measured at baseline – measured at follow-up) / scanning interval; β = standardized regression coefficients.

4. Discussion

To our knowledge, this is the first study to investigate longitudinal changes in brain structure and their relationship with SPS in Shidu parents without any psychiatric disorders. The results showed that Shidu parents had a smaller surface area in the left inferior parietal cortex than healthy controls at baseline and follow-up. Moreover, Shidu parents showed slower rates of cortical thinning and surface area loss in several regions than healthy controls over time, and slower rates of cortical thinning in the left insula, superior frontal cortex, and superior temporal cortex were associated with greater reductions in SPS scores.

The inferior parietal cortex is involved in processing various sensory, perceptual, and cognitive functions (Tabassi Mofrad and Schiller, 2022). A previous functional MRI study reported that individuals exposed to early life stress showed decreased regional homogeneity in the inferior parietal cortex and decreased resting-state functional connectivity between the inferior parietal cortex and the right precuneus, posterior cingulate, and left fusiform gyrus (Philip et al., 2013). This study found that surface area abnormalities in the inferior parietal cortex of Shidu parents persisted over time in the absence of treatment, supporting previous functional findings that trauma exposure had a long-term traumatic impact on the inferior parietal cortex. However, this study found no association between SPS and the surface area of the inferior parietal cortex. Given that Shidu trauma is extremely painful, damage to brain structures may be related to the trauma itself, as revealed in previous studies of our team (Luo et al., 2019; Luo et al., 2016; Luo et al. 2017).

Greater cortical thickness in adults is commonly interpreted to reflect better cortical integrity (Yuan and Raz, 2014; Burzynska et al., 2012). Moreover, greater cortical thickness in brain regions known to be involved in emotion processing, such as the temporal region, is associated with better stress resilience in adults (Kahl et al., 2020). In this study, Shidu parents showed slower rates of cortical thinning over time in multiple regions, including the prefrontal, temporal, and insular regions, than healthy controls. These regions have been shown to be involved in emotional processing of stressful events (Jeong et al., 2021; Nilsen et al., 2016; Giotakos, 2020). Thus, we speculated that slower cortical thinning in Shidu parents, which represents cortical expansion after accounting for healthy aging effects, may reflect a structural improvement process, resilience, or compensatory mechanism in the face of external stress.

The superior temporal cortex plays a key role in speech perception and auditory processing (Bhaya-Grossman & Chang, 2022). Abnormalities in this region have been implicated in the pathogenesis of PTSD (Li et al., 2022; De Bellis et al., 2002; Wang et al., 2021; Geuze et al., 2008). One of the core symptoms of PTSD is re-experiencing, which is characterized by intrusions of memories of traumatic events, such as somatosensory intrusions, visual memories of traumatic events, and recurrent auditory memories (Li et al., 2022; Ehlers et al., 2002). A previous study found that up to 80% of the elderly experience intrusive auditory memories after bereavement (Grimby, 1993). In this study, a slower rate of cortical thinning in the left superior temporal cortex was associated with a greater reduction in trauma re-experiencing symptoms score in Shidu parents, which suggests that the expansion of the temporal region may contribute to improving re-experiencing symptoms and that such symptoms are more likely to point to intrusive auditory memories.

The insula is involved in cognitive control and emotional recognition (Couto et al., 2013). A meta-analysis showed reduced insular activity in patients with PTSD (Koch et al., 2016), and several studies have also reported that insular volume (Herringa et al., 2012) and cortical thickness (Jeong et al., 2019) are negatively correlated with the severity of PTSD symptoms. A recent study found that insular volume decreased over time, and symptoms worsened in PTSD patients with no response to psychotherapy (Zantvoord et al., 2021). This study adds to the literature by discovering that a slower rate of cortical thinning in the left insula was associated with a greater reduction in avoidance symptoms score in Shidu parents, suggesting that the expansion of the insular region may be a protective or compensatory mechanism in the face of stress and may effectively reduce avoidance responses by actively responding to trauma-related emotional stimuli.

The superior frontal cortex is a part of the prefrontal cortex, which is implicated in emotional regulation (Goldberg et al., 2006). A meta-analysis reported thinning of the superior frontal cortex in depressive patients (Li et al., 2020). Another study of Vietnamese ex-political detainees found that cortical thickness in the superior frontal cortex was negatively associated with the severity of depressive symptoms (Mollica et al., 2009). A longitudinal functional MRI study showed that activation of the superior frontal cortex was positively correlated with the resilience of depressive symptoms after childhood maltreatment (Rodman et al., 2019). This study found that a slower rate of cortical thinning in the left superior frontal cortex was associated with a greater reduction in depression symptoms score in Shidu parents, which was consistent with previous studies. This suggests that the expansion of the superior frontal cortex might have helped to reduce depression symptoms and likely reflected resilience achieved in Shidu parents by potentially enhancing emotional regulation in the prefrontal region. Thus, the prefrontal region may be a potential target for improving depressive emotions in Shidu parents.

However, the present study also has several limitations. First, the scanning interval was 0.29 years shorter in Shidu parents than in healthy controls during longitudinal follow-up, but longitudinal changes in brain structure were calculated as an annual rate of change (Abé et al., 2022) to minimize its potential effects. Additionally, the study only involved Shidu trauma, and the results should be generalized with caution to other types of traumatic events or populations. Finally, this study did not include Shidu parents with psychiatric disorders, which limits an in-depth analysis of trauma psychopathology. Further refinement will require larger sample sizes.

In conclusion, the findings of this study revealed that Shidu trauma-induced structural abnormalities in the inferior parietal cortex may persist over time in the absence of treatment, and may be independent of the severity of psychiatric symptoms. Moreover, the expansion of prefrontal, temporal, and insular cortex implicated in emotional regulation may contribute to improvements in psychiatric symptoms, which can provide promising targets for future interventions for Shidu parents.

Acknowledgements

The authors thank all subjects for their willingness to participate in this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to their containment of information that could compromise the privacy of research participants.

Additional information

Funding

This work was supported by Natural Science Foundation of Jiangsu Province: [Grant Number BK20221554]; Foundation of The Commission of Health and Family Planning of Jiangsu Province: [Grant Number ZD2022002]; Foundation of Translational Medicine of Wuxi: [Grant Number ZH202108]; Top Talent Support Program for young and middle-aged people of Wuxi Health Committee: [Grant Number BJ2020109]; Natural Science Foundation of Jiangsu Province: [Grant Number BE2022705].