AUTHORS
Daniel Sop, PhD, Liangsuo Ma, PhD2, Wally Smith I, MD3
AFFILIATIONS
1 Virginia Commonwealth University, Richmond, Virginia, United States
2 Virginia Commonwealth University, Richmond, Virginia, United States
3 Virginia Commonwealth University, Richmond, Virginia, United States
ABSTRACT
Background
Pain in sickle cell disease (SCD) is increasingly recognized as a disorder involving altered central nervous system processing in addition to recurrent vaso-occlusion. Although cerebral hyperperfusion and functional brain abnormalities have each been reported in SCD, little is known about how regional cerebral blood flow relates to the organization of directed brain networks involved in pain processing. We hypothesized that resting cerebral perfusion is associated with the topology of directed pain-network communication and contributes to the heterogeneity of pain phenotypes observed in adults with SCD.
Methods
Eight participants (6 adults with SCD and 2 healthy controls) underwent multimodal MRI including arterial spin labeling (ASL) to quantify regional cerebral perfusion and resting-state functional magnetic resonance imaging (fMRI). Resting-state fMRI data were analyzed using spectral dynamic causal modeling (DCM) to estimate directed effective connectivity. For each participant, 144 effective connectivity parameters were estimated from a DCM comprising 14 pain-related brain regions as network nodes. Graph-theoretical measures, including node strength, PageRank, eigenvector centrality, betweenness centrality, network density, reciprocity, and global efficiency, were calculated after excluding self-connections and retaining effective connections with posterior probabilities ≥0.95. Regional perfusion measurements from the medial prefrontal cortex (mPFC), anterior cingulate cortex (ACC), and precuneus were related to network topology, pain sensitivity, cognition, and hemoglobin concentration using Spearman rank correlations. Given the pilot nature of the study, analyses emphasized effect sizes, consistency of observed relationships, and biological plausibility rather than formal hypothesis testing.
Results
Regional cerebral perfusion demonstrated consistent moderate-to-large relationships with directed network organization. The strongest perfusion-network association was observed between ACC perfusion and PageRank centrality (ρ = −0.905, P = 0.002), while precuneus perfusion demonstrated a positive association with signed incoming network strength (ρ = 0.762, P = 0.028), indicating that resting cerebral blood flow was associated with the organizational influence of key pain-processing regions.
Exploratory analyses demonstrated a consistent pattern of moderate-to-large effect sizes across pain sensitivity, cognition, and hematologic measures, including 201 associations with |ρ| ≥ 0.50, 60 with |ρ| ≥ 0.70, and 9 with |ρ| ≥ 0.90. The most consistent findings localized to the right thalamus, where multiple independent network metrics demonstrated strong inverse relationships with pain sensitivity, including betweenness centrality with Pain Sensitivity Questionnaire (PSQ) Total (ρ = −0.970, P < 0.001), PSQ Minor (ρ = −0.970), PSQ Moderate (ρ = −0.964), and PainDetect (ρ = −0.886). Additional large effect sizes linked left insular eigenvector centrality with pain sensitivity (ρ = 0.839–0.905), right inferior frontal gyrus network influence with hemoglobin concentration (ρ = −0.850), and right insular betweenness centrality with cognitive flexibility (ρ = 0.825). Collectively, these findings demonstrated a coherent pattern linking regional cerebral perfusion, directed pain-network organization, pain sensitivity, cognitive performance, and hematologic status.
Conclusions
This pilot study demonstrates the feasibility of integrating ASL perfusion imaging, spectral DCM-derived effective connectivity, and graph-theoretical network analysis to characterize the hemodynamic organization of directed pain networks in adults with SCD. Rather than examining isolated brain regions or individual connections, this multimodal framework characterizes how regional cerebral perfusion relates to the overall architecture of pain-processing networks. The convergence of moderate-to-large effect sizes across independent imaging and clinical domains supports the concept that altered brain network organization may contribute to pain heterogeneity in SCD. These findings provide a foundation for larger multimodal neuroimaging studies aimed at developing brain-based biomarkers for precision pain phenotyping, patient stratification, and individualized pain management in SCD.
DISCLOSURES
Daniel Sop, PhD: No relevant conflicts of interest to declare.