Effective Connectivity and Clinical Correlates in Chronic Obstructive Pulmonary Disease

Cognitive impairment is a substantial comorbidity in chronic obstructive pulmonary disease (COPD); however, the neural mechanisms underlying this association remain poorly understood. This study aimed to assess alterations in effective connectivity (EC) in patients with COPD and investigate their relationships with clinical measures.

Resting-state functional magnetic resonance imaging (RS-fMRI) was performed in 38 patients with COPD and 30 healthy controls. EC was estimated using vector autoregressive (VAR) modelling combined with Granger causality analysis across 26 prespecified brain regions of interest derived from our previous functional connectivity study of the same cohort. Between-group differences in directed connectivity were assessed using Welch's t-tests. Associations between EC measures showing between-group differences and clinical measures were evaluated using age-adjusted partial Pearson correlations.

At the nominal two-sided p < 0.05 level, 37 directed connections showed between-group differences in effective connectivity, and the mean vector autoregressive coefficients for 35 of these connections lay on opposite sides of zero between the COPD and control groups. Among the connections showing between-group differences, age-adjusted associations in patients with COPD involved pulmonary function, systemic oxygenation, cognition, affect, and health status. These associations involved distributed cortical, subcortical, and cerebellar regions, with several directed connections showing convergence across multiple clinical domains.

Resting-state VAR-based EC revealed distributed between-group differences in direction-sensitive BOLD temporal organization in COPD. Several of these EC measures were associated with pulmonary function, SpO₂, respiratory and COPD symptom burden, cognition, affect, and health status after age adjustment. These findings extend the functional connectivity profile of COPD by linking directed temporal relationships to its multidimensional clinical burden.