Altered brain function in persistent postural perceptual dizziness: A study on resting state functional connectivity

• Published in: Human brain mapping Vol. 39; no. 8; pp. 3340 - 3353
• Main Authors: Lee, Jin‐Ok, Lee, Eek‐Sung, Kim, Ji‐Soo, Lee, Young‐Beom, Jeong, Yong, Choi, Byung Se, Kim, Jae‐Hyoung, Staab, Jeffrey P.
• Format: Journal Article
• Language: English
• Published: United States John Wiley & Sons, Inc 01.08.2018; John Wiley and Sons Inc
• Subjects: Algorithms; Anxiety; Area Under Curve; Artificial intelligence; Biomarkers; Brain; Brain - diagnostic imaging; Brain - physiopathology; Brain Mapping; Cerebellum; Cognition; Comorbidity; Correlation analysis; Cortex (frontal); Cortex (insular); Cortex (parietal); Disability Evaluation; Dizziness; Dizziness - diagnostic imaging; Dizziness - epidemiology; Dizziness - physiopathology; Female; Frontal gyrus; functional connectivity; Functional magnetic resonance imaging; Hippocampus; Humans; Image processing; Information processing; Learning algorithms; Machine learning; Magnetic Resonance Imaging; Male; Medical imaging; Mental depression; Mental Disorders - diagnostic imaging; Mental Disorders - epidemiology; Mental Disorders - physiopathology; Middle Aged; Neural networks; Neural Pathways - diagnostic imaging; Neural Pathways - physiopathology; Neuroimaging; Nucleus accumbens; Occipital lobe; Occipital lobes; Patients; persistent postural perceptual dizziness; Posture; Rest; resting state functional imaging; ROC Curve; Sensory integration; Support Vector Machine; Temporal cortex; vertigo; Vestibular system; Visual cortex; functional connectivity; support vector machine; persistent postural perceptual dizziness; vertigo; resting state functional imaging
• ISSN: 1065-9471; 1097-0193; 1097-0193
• DOI: 10.1002/hbm.24080

This study used resting state functional magnetic resonance imaging (rsfMRI) to investigate whole brain networks in patients with persistent postural perceptual dizziness (PPPD). We compared rsfMRI data from 38 patients with PPPD and 38 healthy controls using whole brain and region of interest analyses. We examined correlations among connectivity and clinical variables and tested the ability of a machine learning algorithm to classify subjects using rsfMRI results. Patients with PPPD showed: (a) increased connectivity of subcallosal cortex with left superior lateral occipital cortex and left middle frontal gyrus, (b) decreased connectivity of left hippocampus with bilateral central opercular cortices, left posterior opercular cortex, right insular cortex and cerebellum, and (c) decreased connectivity between right nucleus accumbens and anterior left temporal fusiform cortex. After controlling for anxiety and depression as covariates, patients with PPPD still showed decreased connectivity between left hippocampus and right inferior frontal gyrus, bilateral temporal lobes, bilateral insular cortices, bilateral central opercular cortex, left parietal opercular cortex, bilateral occipital lobes and cerebellum (bilateral lobules VI and V, and left I–IV). Dizziness handicap, anxiety, and depression correlated with connectivity in clinically meaningful brain regions. The machine learning algorithm correctly classified patients and controls with a sensitivity of 78.4%, specificity of 76.9%, and area under the curve = 0.88 using 11 connectivity parameters. Patients with PPPD showed reduced connectivity among the areas involved in multisensory vestibular processing and spatial cognition, but increased connectivity in networks linking visual and emotional processing. Connectivity patterns may become an imaging biomarker of PPPD.