During the acute phase of Ebola virus disease, neurological symptoms including headache, encephalitis, seizures, and cognitive difficulties are often reported, and there is growing evidence for long-lasting neurological and cognitive compromise in Ebola survivors 1-3. To better understand the relevant neurobiology, it would be valuable to have a non-human primate model amenable to serial evaluation with translational non-invasive brain imaging methods. Such studies have heretofore not been possible given the need for high-level acute containment of infected animals. NIAID has recently established the Integrated Research Facility (IRF) which provides a unique opportunity for performing non-invasive brain imaging (MRI, CT, PET, and SPECT) within a BSL-4 environment. As a first step towards the development of a NHP model to look at neurobiological deficits in Ebola survivors, we examined the acute impact of Ebola on brain resting-state functional connectivity.The main goal of this study was to assess the impact of Ebola on brain function during the acute phase of the disease.

Pre and post-Ebola exposure MRI data were collected from four isoflurane sedated rhesus macaques. Animals were imaged using a Philips Achieva 3 Tesla MR clinical scanner equipped with a Head SENSE Coil. Multi-slice, FFE, EPI-based resting state fMRI data were collected over the course of 7 minutes, with an in-plane resolution of 1.5 x 1.5 mm2 and a slice thickness of 1.5 mm. Other imaging parameters included a TR/TE of 2070/25 ms, NSA = 1, 100 dynamic scans, fat saturation using SPIR, EPI factor = 31, and a FOV = 96 mm x 96 mm. Data were collected a few days prior to Ebola infection and 7-10 days after exposure. Animals were infected with Ebola (Makona strain) through an intramuscular route, with a target dose of 1,000 TCID50.

Pre-processing of the resting-state fMRI data was performed using AFNI4. Following rigid body head motion correction, extracted brain data underwent spatial normalization with respect to a macaque atlas5. Subsequently, data were smoothed using 3mm FWHM Gaussian filter. Prior to performing group independent component analysis (GICA), the time course of each voxel was normalized with respect to its variance. Using the GIFT Toolbox6 a single GICA was performed to extract 20 independent spatial sources, using all of the data from both pre and post-exposures session with each of the 4 macaques. Subject-specific component maps and associated time courses were obtained using back projection. Artifact components and components of biological interest were separated by (1) evaluating the dynamic range and amplitude of low frequency power in the time course spectrum of each component7and (2) evaluation of the spatial profile of the ICA-extracted intrinsic connectivity networks (ICNs) with respect to the profiles of components reported in prior literature8,9.

Six ICNs of interest were identified. These included the default-mode network, a pre-frontal network, auditory and sensorimotor networks, and two visual networks (primary and association). For three of these (default mode, pre-frontal, and primary visual), Ebola infection had minimal impact on functional connectivity. In contrast, the auditory, sensorimotor, and association visual networks showed marked, Ebola-related reductions in functional connectivity (see figure 1). Concurrent T1/T2 structural imaging did not reveal any gross lesions, although some generalized mild atrophy and ventricular dilation was present for some animals. Susceptibility weighted imaging revealed diffuse subarachnoid hemorrhage and venous congestion.The most likely explanation for network-specific disruptions in connectivity is a selective vulnerability in the disrupted networks with respect to (1) a general Ebola-related reduction in hemodynamic compliance [animals were generally at the time of the second scan), (2) local mass/pressure-effects from the hemorrhagic blood products leaked into sub-arachnoid regions around blood vessels, and/or (3) an Ebola triggered neuro-inflammatory response. Future studies, to include more time points and additional measures of regional blood flow and hemodynamic response, will help to clarify these mechanisms of action. Functional brain imaging of Ebola infected NHP is possible, even when animals are maintained in a BSL-4 environment. Data indicate that some ICNs are more vulnerable to perturbation than others. This study sets the stage for future work to explore the neurobiological impact of Ebola on survivors.