Resting-State Effective Connectivity in Sickle Cell Disease Compared With Healthy Controls
Daniel M Sop, Yue M Zhang, Wally R SmithAbstract
Background
Pain in sickle cell disease (SCD) is not fully explained by peripheral pathophysiology alone, as many patients experience persistent, debilitating pain even outside periods of acute vaso-occlusion. Neuroimaging studies in SCD increasingly suggest altered communication across salience, affective, and executive control networks; however, most prior work has relied on correlational functional connectivity approaches, which cannot distinguish directional influences or separate top-down from bottom-up neural control. Growing evidence from chronic pain populations indicates that dysfunction in descending pain-modulatory circuits plays a critical role in pain persistence and sensitization. Despite this, directional neural communication within canonical pain-modulation pathways—including the insula, anterior cingulate cortex (ACC), medial prefrontal cortex (mPFC), thalamus, and periaqueductal gray (PAG)—has not been systematically examined in SCD.
Methods
Adults with SCD (n = 6) and age-matched healthy controls (n = 6) underwent resting-state fMRI scan on a Philips 3T MRI system (multiband factor=3; repetition time=1.625 seconds). Fourteen pain-related brain regions were defined using Neurosynth pain meta-analysis mapped to the AAL3 atlas, with literature-guided inclusion of the PAG. Fully connected dynamic causal models were estimated at the individual level, followed by Bayesian group-level inference. Directional connectivity effects with posterior probability greater than 0.95 were considered reliable.
Results
Across the 182 modeled directed pathways, 11 pathways (6.0%) demonstrated reliable group-level differences meeting the Bayesian threshold (PP > 0.95; Bayes factor > 20). Compared with healthy controls, individuals with SCD showed a 1.8-fold increase in effective connectivity from the right insula to the PAG, representing the strongest and most reliable group difference (PP ≈ 0.98). Reduced PAG self-inhibition was also observed in SCD, consistent with diminished intrinsic inhibitory tone within a key descending modulatory node. Additional reliable alterations were identified in ACC-to-PAG and mPFC-to-insula pathways (PP > 0.95), reflecting disrupted top-down appraisal and affective regulation. Importantly, 8 of the 11 altered pathways (73%) involved salience or prefrontal inputs to brainstem or descending control regions, indicating focused reorganization within pain-modulatory circuitry rather than diffuse network effects. No comparable directional connectivity alterations were observed in healthy controls.
Conclusions
This pilot study provides the first quantitative evidence of altered directional pain-circuit dynamics in adults with SCD at rest. The observed pattern—characterized by amplified salience-to-brainstem signaling, reduced PAG inhibitory control, and disrupted prefrontal–cingulate regulation—supports a mechanistic model of excessive bottom-up nociceptive drive coupled with impaired endogenous pain inhibition. These statistically reliable, circuit-specific alterations highlight the feasibility of effective connectivity as a biomarker framework for stratifying pain phenotypes and targeting descending modulatory pathways in SCD.