Comprehensive examination of resting state fMRI connectomics yields new insights into brain function deficits in Gulf War illness after accounting for heterogeneity in brain impairment across the ill veteran population

Abstract

An estimated 200,000 veterans (up to 32% of those deployed) of the 1991 Gulf War (GW) suffer from GW illness (GWI), an incompletely understood chronic medical condition, characterized by multiple symptoms indicative of brain function deficits in various domains. Epidemiologic and animal studies have associated GWI with exposure to neurotoxic chemicals such as nerve agents, organophosphate pesticides and pyridostigmine bromide. One factor that hampers mechanistic investigations into GWI is that there is considerable heterogeneity in brain impairments across the ill GW veteran population. This could reflect the underlying heterogeneity in both exposure to neurotoxic substances, as well as genetic predisposition or resistance to neurotoxicity. Only one of the validated case definitions, the Haley GWI criteria addresses this heterogeneity. It does so by breaking down GWI into three main syndrome variants (GWS1, GWS2, and GWS3) based on factor analysis of symptoms presented by GWI veterans. Resting state fMRI (rsfMRI) is a uniquely useful brain imaging technique in that in a 10-min fMRI scan it can probe numerous brain function domains simultaneously. In this study, we employed a connectomics approach and machine learning on rsfMRI data from a cohort of GW veterans to extract neuroimaging biomarkers specific to each of the three Haley GWI syndromes. Our results revealed a number of new insights into brain function impairment specific to each syndrome group. The findings indicate that these deficits may by and large be driven by brain mechanisms. We also found that pooling the data of all three syndromes in GWI group, as is done by commonly employed case definitions of GWI resulted in failure to detect the fMRI signatures of a lot of these brain impairments.