Changes in resting-state functional connectivity in neuropsychiatric lupus: A dynamic approach based on recurrence quantification analysis. (February 2022)

• Record Type: Journal Article
• Authors: Pentari, Anastasia; Tzagkarakis, George; Tsakalides, Panagiotis; Simos, Panagiotis; Bertsias, George; Kavroulakis, Eleftherios; Marias, Kostas; Simos, Nicholas J.; Papadaki, Efrosini
• Is Part Of: Biomedical signal processing and control. Volume 72(2022)Part A
• Publication Date: 2022-02
• Subjects: Dynamic functional connectivity -- Recurrence quantification analysis -- Neuropsychiatric systemic lupus erythematosus; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2021.103285
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Investigation of the cross recurrence quantification analysis (CRQA). Application on 16 brain regions which consist the visuomotor network. Comparison with static functional connectivity (FC)-Pearson's correlation. Comparison with temporal-based dFC alternative approach. CRQA proved to be more efficient in separating NPSLE patients from HC volunteers. Abstract: There is growing interest in dynamic approaches to functional brain connectivity (FC), and their potential applications in understanding atypical brain function. In this study, we assess the relative sensitivity of cross recurrence quantification analysis CRQA) to identify aberrant FC in patients with neuropsychiatric systemic lupus erythematosus (NPSLE) in comparison with conventional static and dynamic bivariate FC measures, as well as univariate (nodal) RQA. This technique was applied to resting-state fMRI data obtained from 45 NPSLE patients and 35 healthy volunteers (HC). Cross recurrence plots were computed for all pairs of 16 frontoparietal brain regions known to be critically involved in visuomotor control and suspected to show hemodynamic disturbance in NPSLE. Multivariate group comparisons revealed that the combination of six CRQA measures differentiated the two groups with large effect sizes ( . 214 > η 2 > . 061 ) in 40 out of the 120 region pairs. The majority of brain regions forming these pairs also showed group differences on nodal RQA indices ( . 146 > η 2 > . 09 ) Overall, larger values were found in NPSLE patients vs. HC with the exception of FC formed by the paracentral lobule. Determinism within five pairs of right-hemisphere sensorimotor regions (paracentral lobule, primary somatosensory, primary motor, and supplementary motor areas), correlated positively with visuomotor performance among NPSLE patients ( p < . 001 ). By comparison, group differences on static FC displayed large effect sizes in only 4 of the 120 region pairs ( . 126 > η 2 > . 061 ), none of which correlated significantly with visuomotor performance. Indices derived from dynamic, temporal-based FC analyses displayed large effect sizes in 11 / 120 region pairs ( . 11 > η 2 > . 063 ). These findings further support the importance of feature-based dFC in advancing current knowledge on correlates of cognitive dysfunction in a clinically challenging disorder, such as NPSLE.