Differences in Brain Structure and Functional Connectivity During Memory Processing in Posttraumatic Stress Disorder
Article information
Abstract
Objective
This study aimed to investigate the association between changes in brain volumes and memory task-based functional connectivity (FC) patterns in patients with posttraumatic stress disorder (PTSD) compared to healthy controls (HCs).
Methods
This study employed voxel-based morphometry and task-based functional magnetic resonance imaging using memory task to examine alterations in brain volume density and FC within and between relevant brain regions involved in memory processes such as encoding and retrieval.
Results
PTSD patients exhibited increased brain volumes in the right cerebellum (Cb) in the gray matter (GM) region compared to HCs. Analysis of FC patterns revealed that both groups exhibited similar connectivity patterns, but PTSD patients displayed higher and more extensive interregional connectivity relative to HCs. During the memory encoding process, PTSD patients exhibited higher FC linked to the Cb in the right postcentral gyrus, left precentral gyrus (PrCG), left orbito frontal gyrus (OFG), and superior frontal gyrus, while demonstrating lower connectivity in the right hippocampus. In the memory retrieval period, PTSD patients showed higher FC in the right middle frontal gyrus, left middle occipital gyrus, left OFG, right inferior parietal gyrus, and right PrCG, along with lower FC in the right/left inferior frontal gyrus and right superior parietal gyrus. Additionally, a negative correlation was observed between PTSD symptom severity and local GM volumes in the right Cb. However, no significant correlation was found with FC.
Conclusion
Our findings provide insights into the specific alterations in brain structure and connectivity associated with PTSD, contributing to a better understanding of the underlying neural mechanisms of PTSD.
INTRODUCTION
Posttraumatic stress disorder (PTSD) is a debilitating psychiatric condition that can develop in individuals who have experienced or witnessed traumatic events. PTSD is characterized by a range of symptoms, including intrusive memories, hyperarousal, avoidance behaviors, and negative alterations in cognition and mood [1]. Among these symptoms, disturbances in memory processes have been recognized as a significant feature of PTSD [2].
Memory dysfunction in PTSD is commonly observed in both the encoding and retrieval phases of memory, resulting in impairments in the formation and retrieval of traumatic and non-traumatic memories [3]. These memory disturbances contribute to the persistent re-experiencing of traumatic events, difficulty in emotional regulation, and the maintenance of PTSD symptoms [4].
In recent years, neuroimaging techniques have provided valuable insights into the neural underpinnings of PTSD-related memory dysfunction [2]. Among these techniques, functional magnetic resonance imaging (fMRI) has emerged as a powerful tool to investigate the functional connectivity (FC) of brain networks during memory tasks [5]. FC refers to the temporal synchronization and coordination of brain regions during task performance or rest, reflecting the communication and integration of neural networks.
Previous studies [6-9] have revealed alterations in FC patterns associated with memory process in individuals with PTSD, implicating disrupted interactions between key brain regions involved in memory processes. These alterations involve regions such as the amygdala, hippocampus (Hipp), prefrontal cortex, and the default mode network, which are crucial for memory encoding, consolidation, and retrieval [2,6-9]. Understanding the specific alterations in memory task-based FC in PTSD patients can shed light on the underlying neural mechanisms of memory dysfunction and potentially guide the development of targeted interventions. Furthermore, such investigations may also contribute to the identification of potential biomarkers for early detection and monitoring of treatment response in PTSD.
Brain function is a complex phenomenon influenced by various factors. Neuronal cell density refers to the number of neurons present in a specific area of the brain. The brain consists of billions of neurons that communicate with each other through intricate networks. Also, there are many neuroimaging studies [10-12] explored association between changes of brain volumes and FC.
The present study aims to investigate association between changes of brain volumes and memory task-based FC patterns in PTSD patients using voxel-based morphometry (VBM) and fMRI by employing memory tasks that engage different memory processes, such as encoding and retrieval. We aim to elucidate the specific alterations in FC within and between relevant brain regions affected by abnormal brain volume density in patients with PTSD compared to healthy controls (HCs).
Findings from this study have the potential to advance our understanding of the neural correlates between brain volume difference and its effect to memory dysfunction in PTSD and may inform the development of novel therapeutic approaches for alleviating memory-related symptoms in this population. Furthermore, this research may contribute to the broader field of neuroimaging and help uncover mechanisms underlying memory processes in healthy individuals as well.
In summary, investigating brain volume changes and memory task-based FCs in PTSD patients has the potential to uncover critical insights into the neuronal cell defects and its association to memory dysfunction derived from PTSD symptomatology. The findings may pave the way for innovative interventions aimed at improving memory function and overall well-being in individuals affected by PTSD.
METHODS
Subjects
We recruited seventeen patients and 17 control subjects. However, due to poor MRI data quality, 8 individuals were excluded from the analysis. Finally, thirteen PTSD patients (mean age, 27.92±10.09 years) and 13 HC (mean age, 29.08±9.40 years) (Mann-Whitney U test, p=0.687) diagnosed by a psychiatrist based on the Diagnostic and Statistical Manual of Mental Disorders 5 (DSM-5) criteria were included in this study (Table 1). All participants were right-handed (chi-square test, p>0.999) and took the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) [13]. The CAPS-5 is a validated tool utilized in the assessment of the severity, validity, and improvement of PTSD symptoms as outlined by the DSM-5. Each PTSD symptom is evaluated using separate 5-point rating scales for frequency and intensity, ranging from 0 to 4. Participants with CAPS-5 scores below 40 were excluded from the subsequent analysis. Before undergoing MRI, all volunteers received a thorough explanation of all experimental procedures and each provided written informed consent. Participants in both groups showed no abnormalities on physical and neurological examinations. Also, all the experimental procedures and methods were performed in accordance with the relevant guidelines and regulations approved by the IRB-JBUH (Jeonbuk National University Hospital IRB number 2019-07-016).
Demographics between patients PTSD and healthy controls
In addition, clinical characteristics of the PTSD group were assessed. The mean duration of illness was 49.7 months (range, 6–125 months). Trauma types varied across individuals and included childhood peer bullying (n=6), sexual assault (n=1), physical abuse by family members (n=1), motor vehicle accidents (n=3), military-related assault (n=1), and the traumatic loss of a close acquaintance (n=1). All participants were prescribed antidepressants, including escitalopram (n=4, 10–20 mg), sertraline (n=4, 50–100 mg), desvenlafaxine (n=3, 50– 100 mg), vortioxetine (n=1, 10 mg), fluvoxamine (n=1, 100 mg), and mirtazapine (n=1, 15 mg). Concomitant use of anxiolytics, hypnotics, and antipsychotics was also observed, including etizolam (n=4, 0.25–0.75 mg), zolpidem (n=2, 10–12.5 mg), aripiprazole (n=2, 2–4 mg), and others (e.g., trazodone [n=1, 25 mg], clonazepam [n=1, 0.5 mg], alprazolam [n=1, 0.25 mg]).
Inclusion criteria
1) Participants who met DSM-5 criteria for PTSD.
2) Age of participants were between 20 and 60 years.
3) All participants have an education above middle school (more than 9 years: middle school in Korea).
4) Right-handed.
Exclusion criteria
1) History of head trauma.
2) History of cardiovascular or endocrine disease.
3) Psychiatric illness other than PTSD.
4) Presence of magnetically active object in the body.
Data acquisition
All participants were subjected to neuroimaging using a 3 tesla Magnetom Verio MR Scanner (Siemens Medical Solutions) equipped with an 8-channel birdcage head coil. Highresolution T1-weighted images were obtained through a threedimensional magnetization prepared rapid acquisition gradient echo (3D MP-RAGE) sequence, employing the following acquisition parameters: repetition time (TR) of 1,900 ms, echo time (TE) of 2.35 ms, a field-of-view (FOV) of 22×22 cm2, matrix size of 256×256, one excitation (NEX), and a slice thickness of 1 mm. Functional images were acquired utilizing a gradient-echo echo planar imaging technique, with the subsequent acquisition settings: TR of 2,000 ms, TE of 30 ms, flip angle of 90°, FOV of 22×22 cm2, matrix size of 64×64, one NEX, 25 slices, and a slice thickness of 5 mm without a slice gap.
Paradigm for memory task
Memory task was conducted utilizing two-syllable words with negative emotional content (Figure 1). The brain activation paradigm followed a specific cycle: a resting phase lasting 14 seconds, followed by the first encoding phase lasting 18 seconds, another resting phase of 14 seconds, the first retrieval phase lasting 18 seconds, a subsequent resting phase of 14 seconds, the second encoding phase lasting 18 seconds, another resting phase of 14 seconds, the second retrieval phase lasting 18 seconds, and a final resting phase of 14 seconds. Throughout the rest periods, two fixation crosses were displayed. A set of six different two-syllable words was displayed for 3 seconds each during the encoding phase. Subsequently, during the retrieval phase, the two-syllable words presented during the encoding phase, but without the first consonant, were presented. Then, participants were presented with either new words or the previously presented words from the “encoding” phase. Participants were instructed to press a button if the presented word was recognized as an old word. All word stimuli were presented on a computer monitor employing Superlab Pro software (Cedrus Co.).
Activation paradigm of memory task with unpleasant words. In each encoding period, six different two-syllable words, each appearing once, were sequentially presented. In the following retrieval period, the old words presented in encoding period and new words, without the first consonant, were randomly presented.
Brain volume change analysis
The anatomical images were post-processed using statistical parametric mapping (SPM12) software for VBM analysis following the procedure described in our previous study [14]. Alternating gray matter (GM) and white matter (WM) volumes in both groups were assessed with an independent two-sample t-test with multiple comparisons using family-wise error (FWE) at p<0.05 and cluster size >100 voxels.
Task-based FC analysis
The FC analysis was conducted using the CONN-fMRI FC toolbox (ver. 21a) in conjunction with the SPM12 software. Preprocessing of the data involved several steps. Firstly, slice timing correction was applied using an interleaved acquisition scheme. Subsequently, image realignment was performed to correct for subject motion, with a threshold of 2 mm used to identify and correct for motion artifacts. Field map correction was also applied. The individual T1 images were then co-registered with the functional images. Segmentation of the transformed T1 images into GM, WM, and cerebrospinal fluid was performed using standard SPM tissue probability maps. To reduce spatial noise, the images underwent spatial smoothing using a Gaussian kernel with a full width at half maximum of 8 mm3. For robust detection of functional outliers, a liberal setting was employed, utilizing the 97th percentiles in a normative sample and a global-signal z-value threshold of 9. An anatomical component-based noise correction method (aCompCor) was employed to minimize the impact of noises on the bloodoxygenation-level-dependent (BOLD) signals, including subject motion and identified outlier scans or scrubbing artifacts [15]. Finally, a band-pass filter was applied, restricting the frequency window of the BOLD signals to 0.01–0.1 Hz. Right cerebellum (Cb) (x, y, z=15, -62, -27) observed in VBM analysis was designated as seed region for seed to voxel FC analysis.
Statistical analysis between clinical data and brain volume changes and FC
To analyze the correlation between CAPS-5 scores and different GM volumes and FCs and, Pearson’s correlation coefficient test was performed using the SPSS Statistical software package (version 20.0, IBM Corp.).
RESULTS
Demographic and clinical measurements
Table 1 shows demographic and clinical characteristics of the patients with PTSD and HC. No differences between the two groups were noted in terms of age, sex, handedness and education. The mean scores for CAPS-5 in PTSD patients was 53.92±10.14.
Brain volume difference
There was no significant volume difference in the WM regions between the two groups. However, in the GM regions, patients with PTSD showed increased brain volumes in right Cb (x, y, z=15, -62, -27) when compared to HC (two sample t-test, p-FWE<0.05, cluster size >100 voxel) (Figure 2).
Increased gray matter volumes in patients with posttraumatic stress disorder compared to healthy controls (two sample t-test, uncorr. p<0.001, cluster size >100 voxel). L, left; R, right; Cb, cerebellum (x, y, z=15, -62, -27).
Differences in the FC
The FC patterns derived from the Cb between patients with PTSD and HC are shown in Figure 3 and Table 2. Both groups are characterized by a similar pattern, but have higher and wider interregional connectivity in patients with PTSD than in HC relatively. During memory encoding process, PTSD patients showed relatively higher FCs with the Cb in right the postcentral gyrus (PoCG), left precentral gyrus (PrCG), left orbito frontal gyrus (OFG), and superior frontal gyrus (SFG). On the other hand, there was lower connectivity in right Hipp. In memory retrieval period, there were higher FC in right middle frontal gyrus (MFG), left middle occipital gyrus (MOG), left OFG, right inferior parietal gyrus (IPG), and right PrCG, and lower FC in right/left inferior frontal gyrus (IFG) and right superior parietal gyrus (SPG) in patients with PTSD (one sample t-test, voxel threshold: uncorr. p<0.001, cluster threshold: p-FDR<0.05). In the group difference analysis determined by a two-sample t-test, brain regions including the inferior temporal gyrus (ITG), IPG, SFG, middle cingulate gyrus, and PoCG were identified during the retrieval task only, using threshold-free cluster enhancement with subsequent FWE correction (Figure 4 and Table 3).
Positive connectivity networks associated with right cerebellum showing increased gray matter volumes in posttraumatic stress disorder (PTSD) patients and health controls during encoding (A and B) and retrieval periods (C and D) (one sample t-test, voxel threshold: uncorr. p<0.001, cluster threshold: p-FDR<0.05).
Group difference in localized functional connectivity during implicit memory task
Positive connectivity networks associated with the right cerebellum showing increased gray matter volumes in posttraumatic stress disorder (PTSD) patients and healthy controls during the encoding (A and B) and retrieval periods (C and D), as determined by a two-sample t-test.
Regions showing predominant functional connectivity linked to the right Cb between PTSD patients and healthy controls, as determined by a two-sample t-test
Correlation between PTSD symptoms and GM volume and FC difference
Figure 5 shows a significant negative correlation between CAPS-5 scores and local GM volumes in right Cb in patients with PTSD (Pearson’s correlation (r)=-0.617, p=0.025). However, there was no significant correlation with FC.
The correlation between CAPS-5 scores and brain volume changes of right Cb in patients with PTSD (r2=0.347, p=0.025) (Pearson’s correlation (r)=-0.617). CAPS-5, Clinician-Administered PTSD Scale for DSM-5; Cb, cerebellum; PTSD, posttraumatic stress disorder.
DISCUSSION
The present study aimed to investigate the association between changes in brain volumes and memory task-based FC patterns in patients with PTSD compared to HCs. Our findings revealed several notable results regarding brain volume differences, FC patterns, and the correlation between PTSD symptom severity and brain morphometry. In terms of demographic and clinical characteristics, no significant differences were observed between patients with PTSD and HCs in terms of age, sex, handedness, and education. This suggests that any observed differences in brain measures between the groups are unlikely to be influenced by these demographic factors.
Regarding brain volume differences, our study found increased brain volumes in the right Cb in the GM region of PTSD patients compared to HCs. This finding is consistent with previous research suggesting structural alterations in the Cb in PTSD [16,17]. The Cb is known to play a crucial role in cognitive and emotional processing [18]. Recent research has shown that the Cb plays a role in fear learning and memory [19,20]. Given that PTSD is marked by abnormalities in threat detection and processing [21], this growing body of evidence strongly suggests the involvement of the Cb in the pathophysiology of PTSD. In our result, alterations in cerebellar volume may contribute to the cognitive and emotional dysregulation observed in PTSD [15]. Interestingly, although most previous studies reported decreased cerebellar volumes in PTSD patients, our study found increased GM volumes in the right Cb compared to HCs. A similar pattern of cerebellar enlargement has been reported in adult combat-related PTSD, where volumetric increases were observed particularly in lobules 7b and 8a–b, regions implicated in emotional processing [16]. Consistent with these findings, our results may reflect an early compensatory or maladaptive hypertrophy in response to trauma, which might subsequently be followed by volume reduction associated with symptom progression. Chronic stress-related neuronal loss, impaired neuroplasticity, or neurodegenerative changes could contribute to the observed volume decrease among patients with more severe PTSD symptoms [22]. Similar compensatory-to-degenerative patterns have been described in other stress-related disorders.
Analyzing FC patterns, we found that both the PTSD and HC groups exhibited similar patterns; however, PTSD patients showed higher and more extensive interregional connectivity compared to HCs. During the memory encoding process, PTSD patients displayed higher FC with the Cb in the right PoCG, left PrCG, left OFG, and SFG. This heightened connectivity may indicate compensatory mechanisms or hyperactivation in regions involved in memory encoding [15]. Neurostructural studies have demonstrated that in patients who have experienced trauma, especially those with mild traumatic brain injury, smaller cortical volumes in regions such as the SFG, rostral, and caudal cingulate play a crucial role in predicting the likelihood of developing PTSD symptoms. The SFG is involved in higher cognitive functions, such as working memory, executive control, and attention regulation, which are often impaired in PTSD patients. A study suggests that reduced SFG volume is linked to the difficulty in regulating intrusive memories and cognitive dysfunctions frequently observed in PTSD [23]. Furthermore, Zimmermann et al. [24] observed that the left and right lateral OFG volumes in PTSD patients were significantly smaller than in controls and were negatively correlated with symptom severity. The OFG is associated with emotional regulation, decision-making, and processing of reward-related stimuli. In PTSD, diminished OFG volume may contribute to impaired emotional regulation, which exacerbates the re-experiencing of traumatic events and emotional dysregulation. Taken together, these findings indicate that PTSD patients tend to show smaller cortical volumes in prefrontal regulatory regions, consistent with preclinical work suggesting that prefrontal areas, including the OFG and SFG, play a critical role in the recovery and extinction of threat behaviors. This highlights the importance of these regions in both the emotional and cognitive aspects of PTSD, contributing to a better understanding of how functional and structural changes in the brain contribute to the development and persistence of symptoms. Conversely, lower connectivity was observed in the right Hipp, a region crucial for memory formation [25]. Reduced hippocampal activity in PTSD is also com-mon, and it appears to play an important role in process of fear learning and memory consolidation [26,27]. Our result consistent with previous studies that suggested disrupted hippocampal functioning in PTSD patients during encoding [28,29].
In the memory retrieval period, PTSD patients demonstrated higher FC in the right MFG, left MOG, left OFG, right IPG, and right PrCG. These regions are involved in attention [30], visual processing [31,32], and cognitive control [33,34], suggesting increased engagement of these areas during memory retrieval in PTSD patients. Conversely, lower FC was observed in the right/left IFG and right SPG, regions associated with executive functions and working memory [35,36]. These abnormalities showed in psychiatric disorders. These findings may reflect impaired executive control processes in PTSD during memory retrieval. Interestingly, during this period, PTSD patients exhibited a shift in FC from the SFG and OFG to the MFG and OFG during the transition from memory encoding to retrieval. This shift may indicate alterations in the neural networks associated with PTSD. During memory encoding, the SFG is involved in various cognitive functions such as working memory and executive control [37], while the OFG plays a crucial role in emotional processing and decision-making [38]. In contrast, during memory retrieval, the MFG is engaged in self-referential processing and memory recall [39], and the OFG continues to be involved in emotion- and reward-related processing [38]. A previous study has demonstrated that PTSD patients show increased FC between the MFG and amygdala compared to HCs and trauma-exposed individuals without PTSD [40]. These distinct changes in FC may be critical in the development of PTSD. The altered connectivity patterns observed in our study likely reflect the brain’s attempt to compensate for emotional and cognitive disruptions caused by PTSD. The brain may reorganize FC by recruiting alternative networks to support memory processes and mitigate the effects of the disorder. Interestingly, the brain regions, including the ITG, IPG, SFG, middle cingulate gyrus, and PoCG, were identified during the memory retrieval task only, suggesting that PTSD patients exhibit altered FC in networks related to memory processing, emotional regulation, and sensory integration. These findings align with previous studies indicating that PTSD is associated with impaired top-down control over memory retrieval, heightened sensitivity to trauma-related cues, and dysregulated sensory processing, potentially contributing to core symptoms such as intrusive thoughts, hypervigilance, and emotional dysregulation. Furthermore, we examined the correlation between PTSD symptoms, as measured by CAPS-5 scores, and brain measures. We found a significant negative correlation between symptom severity and local GM volumes in the right Cb, suggesting that as PTSD symptoms worsened, GM volumes in the Cb decreased. Our result is consistent with previous study reported that there was a significant negative correlation between vermal volume and CAPS total score in patients with PTSD [41]. It could be suggested that the Cb is one of critical brain region that involved in cognitive-emotional processing in the PTSD patient. However, no significant correlation was observed between symptom severity and FC. It is important to note that the lack of correlation with FC might be attributed to the small sample size or other factors not accounted for in this study.
Overall, our findings contribute to the growing body of literature on the neurobiological underpinnings of PTSD. The increased brain volumes in the Cb, along with altered FC patterns during memory encoding and retrieval, provide insights into the specific neural alterations associated with PTSD. These findings suggest that PTSD is characterized by both structural and functional changes in brain regions involved in memory processes. Future studies with larger sample sizes and longitudinal designs are warranted to confirm and expand upon these findings.
Understanding the neurobiological mechanisms underlying PTSD can inform the development of targeted interventions and therapies aimed at alleviating symptoms and improving the quality of life for individuals affected by this debilitating disorder.
There are some limitations in our study. First, one limitation of this study is the small sample size, which limited the use of more advanced statistical models such as random or mixed-effects analyses. In addition, covariation analyses between symptom severity and brain measures require larger samples to ensure statistical power, particularly given the complexity of fMRI data. A larger sample would be needed to replicate our findings as well as to investigate further understand neurobiological mechanism of PTSD. Second, the possibility of medication effects on brain structure and FC could not be excluded. Whether brain changes in PTSD are primary itself or secondary to other unknown reasons needs to be further elucidated.
In conclusion, present study found a significant increasement in GM volume in local region of Cb that showed close association with clinical symptom severity (CAPS-5 scores) among PTSD patients, also this brain volume changes led to abnormality of brain FC during memory processing. These findings suggest that PTSD symptom can lead to structural and functional network alterations in key brain regions, providing insights into the neural mechanisms underlying PTSD and potential implications for future treatment research.
Notes
Availability of Data and Material
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
Conflicts of Interest
Jong-Chul Yang, a contributing editor of the Psychiatry Investigation, was not involved in the editorial evaluation or decision to publish this article. All remaining authors have declared no conflicts of interest.
Author Contributions
Conceptualization: all authors. Formal analysis: Shin-Eui Park. Funding acquisition: Jong-Chul Yang. Methodology: Jong-Chul Yang, Shin-Eui Park. Software: Shin-Eui Park. Writing—original draft: all authors. Writing—review & editing: all authors.
Funding Statement
This paper was supported by Fund of Biomedical Research Institute, Jeonbuk National University Hospital.
Acknowledgments
None