Functional Neuroimaging Distinguishes Posttraumatic Stress Disorder from Traumatic Brain Injury in Focused and Large Community Datasets
Level 4 - case-series / case-control
Retrospective case-control and database-derived diagnostic accuracy study.
OpenAlex W1567104108 · doi:10.1371/journal.pone.0129659
What was done
Researchers evaluated the diagnostic accuracy of resting and on-task single photon emission computed tomography (SPECT) in distinguishing posttraumatic stress disorder (PTSD) from traumatic brain injury (TBI). The analysis drew from a multisite database and evaluated two cohorts: Group 1, a matched sample comprising TBI (n=104), PTSD (n=104), comorbid PTSD+TBI (n=73), and healthy controls (n=116); and Group 2, a larger community dataset including TBI (n=7,505), PTSD (n=1,077), PTSD+TBI (n=1,017), and controls without either condition (n=11,147). Region-of-interest (ROI) perfusion data and visual readings were evaluated using binary logistic regression and receiver operating characteristic (ROC) curves, with 10-fold cross-validation applied to Group 2.
What was found
In the matched cohort (Group 1), ROI analysis achieved 100% sensitivity, 100% specificity, and 100% accuracy in separating PTSD, TBI, and comorbid PTSD+TBI, while visual readings achieved 89% or higher across these metrics. For the large community sample (Group 2), accuracy, sensitivity, and specificity were lower than in Group 1, but specific numerical values were not reported in the abstract. Compared with TBI, PTSD exhibited significantly increased perfusion in the limbic regions, cingulum, basal ganglia, insula, thalamus, prefrontal cortex, and temporal lobes.
Why it matters
PTSD and TBI share substantial symptom overlap, complicating clinical diagnosis and management when they co-occur. Demonstrating distinct functional perfusion patterns on SPECT suggests potential utility as an objective biomarker to guide differential diagnosis.
Limits
The study is limited by its retrospective design using a private multisite clinical database. The report of 100% sensitivity and specificity in Group 1 indicates a high risk of model overfitting or spectrum bias. Specific diagnostic metrics and confidence intervals for the primary community cohort (Group 2) are not provided in the abstract, preventing independent assessment of its real-world clinical performance.
Cited by
- supports Dr. Daniel Amen's research team has published more than 80 scientific articles in peer-reviewed journals.