Despite the prevalence of blast injuries in recent overseas conflicts, knowledge of their impact on neural health is lacking. We have recently published work demonstrating differences in functional magnetic resonance imaging (fMRI) connectivity that were specific to close-range blast exposure (CBE), as opposed to other prevalent military-related factors. Here, we replicate this finding in an independent sample of 135 veterans, again finding that CBE, regardless of concussion, is predictive of persistent changes in brain physiology. Although there was weak overlap anatomically, in both samples, the group differences could be described as spreading of anticorrelation. Using the combined sample, we now seek to identify likely mechanisms that could bring about this effect. We compared participants with (n = 116) and without (n = 153) CBE by analyzing two networks through group difference maps and correlation distributions to assess spatially homogenous and heterogeneous effects. As boundaries between positive and negative correlations in fcMRI are determined by noise covariates, we compared analyses with and without global signal regression. We found evidence of widespread altered connectivity that was spatially heterogeneous across participants, and that the role of global signal regression was network dependent. These findings are not consistent with expected results from damaged white matter or impaired neural function. Rather, potential biological interpretations include disrupted cerebral blood flow or impaired neurovascular coupling, which have each been observed in animal models of blast exposure. Further targeted work will be necessary to distinguish the contribution of each of these mechanisms to producing changes in brain function associated with CBE.