PTSD subtypes and their underlying neural biomarkers: a systematic review

Abstract

Posttraumatic stress disorder (PTSD) is a heterogenous disorder with frequent diagnostic comorbidity. Research has deciphered this heterogeneity by identifying PTSD subtypes and their neural biomarkers. This review summarizes current approaches, symptom-based group-level and data-driven approaches, for generating PTSD subtypes, providing an overview of current PTSD subtypes and their neural correlates. Additionally, we systematically assessed studies to evaluate the influence of comorbidity on PTSD subtypes and the predictive utility of biotypes for treatment outcomes. Following the PRISMA guidelines, a systematic search was conducted to identify studies employing brain imaging techniques, including functional magnetic resonance imaging (fMRI), structural MRI, diffusion-weighted imaging (DWI), and electroencephalogram (EEG), to identify biomarkers of PTSD subtypes. Study quality was assessed using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. We included 53 studies, with 44 studies using a symptom-based group-level approach, and nine studies using a data-driven approach. Findings suggest biomarkers across the default-mode network (DMN) and the salience network (SN) throughout multiple subtypes. However, only six studies considered comorbidity, and four studies tested the utility of biotypes in predicting treatment outcomes. These findings highlight the complexity of PTSD's heterogeneity. Although symptom-based and data-driven methods have advanced our understanding of PTSD subtypes, challenges remain in addressing the impact of comorbidities and the limited validation of biotypes. Future studies with larger sample sizes, brain-based data-driven approaches, careful account for comorbidity, and rigorous validation strategies are needed to advance biologically grounded biotypes across mental disorders.

Keywords: biotypes; data-driven approaches; neuroimaging; posttraumatic stress disorder; subtypes.

Figure 1.

This figure illustrates the two approaches to understanding the subtypes of PTSD, including the DSM-based top-down method, symptom-based group analysis, and brain-based data-driven method. The arrow indicates the order of each step in generating subtypes. The orange triangles illustrate the steps to generate PTSD subtypes using a symptom-based group analysis approach. The green triangle on the right describes the steps in creating subtypes using a brain-based data-driven approach.

Figure 2.

Flowchart for study inclusion. Initial literature search was conducted via PubMed, Library of Congress, LISTA, Web of Science Core Collection using keywords (refer to Supplementary Material Methods). Additional literature was added based on reviewing reference lists and a priori knowledge. The PRISMA review methods were used to evaluate the research articles. A final number of 53 studies were included. For additional details, refer to Supplementary Material.

Figure 3.

This figure illustrates the study overview through the following four figures. (a) This graph presents the percentage of available studies using each of the two identified approaches: symptom-based group analysis and brain-based data-driven approach. (b) This graph summarizes the percentage of selected studies that investigated different categories of PTSD subtypes using a top-down approaches; 10 subtypes of PTSD have been under investigation, with the dissociative subtype being the most researched subtype as 40% of studies investigated dissociative subtypes. (c) This graph illustrates the percentage of studies that utilized each type of neuroimaging modality. (d) This graph provides an overview of the percentage of studies that used different imaging feature selection methods. (e) This graph provides an overview of the percentage of studies that investigated each type of trauma. (f) This graph is an overview of the number of literature with the respective sample populations; most studies include a sample size of 66 to 156. (g) This chart provides an overview of the types of tasks used by the task-based fMRI studies. The current review study found a total of 13 fMRI task-based studies; the figure demonstrates that the most used task is the emotional picture task, followed by Go/NoGo task and the Script-driven imagery task. The chart reveals that the tasks used are diverse, each task is used by one to three studies.

Figure 4.

This figure illustrates the neural network alterations within the dissociative subtypes, specifically the difference in the directionality of connectivity within PTSD+DS vs. PTSD-DS. Thick lines indicate two or more studies found consistent directionality of connectivity. The narrow line indicates the directionality of the connectivity is implicated by one study. Arrows illustrate the pattern of between-network connectivity; the square around the network names represents the pattern of within-network connectivity. Increased between-network connectivity is represented by a red line and observed between the following networks: pDMN-brainstem, SN-pDMN, aDMN-Brainstem, ECN-BGN. Increased within-network connectivity observed in aDMN, ECN. Decreased network connectivity is represented by the blue line, observed between the following networks: ECN-Brainstem. Decreased within-network connectivity was found in the following networks: pDMN, SN, Cerebellar network, and Brainstem. An inconsistent network pattern, represented by a grey dotted line, was found between the following networks: SN-Brainstem, Cerebellar-pDMN, ECN-aDMN, ECN-pDMN. Inconsistent within-network connectivity found in BGN. The network is defined as the following brain hubs. pDMN: Posterior DMN, including PCC, precuneus, TPJ. aDMN: anterior DMN, including vmPFC. SN: salience network, including amygdala, insula. ECN: executive control network, including DLPFC, frontal pole, anterior cingulate cortex.