Datensatz

Zusammenfassung

The primate brain is a dynamic system interconnected by temporally correlated functional networks. The structure of this correlated activity depends on the brain’s internal state and on stimulus input. In the absence of external stimulation, functional networks of spontaneous activity,
i.e. the salience network, the executive control network and the default mode network, can be observed. Their origin and function are not well understood, but they could reflect
neural noise within anatomically connected areas or active mechanisms related to perception and awareness. On the other hand, when the brain is being stimulated, a different pattern of activity emerges. Exactly how this patiotemporal transition happens is still unclear.
The objective of this study is to characterize the dynamic changes of BOLD based functional connectivity between resting-state and natural-stimuli-driven networks in the awake monkey brain. Due to its high spatial resolution, BOLD-fMRI is a powerful tool to study large-scale
correlated brain network activity. We used a paradigm containing sequences of movie-clips with different contexts including natural and artificial environments as well as periods devoid of any visual stimulation (resting) in order to identify the global activation patterns reflecting
the interplay between different populations of neurons under these conditions.
For our experiments, two macaque monkeys (Macaca mulatta) were trained in a mock scanner to remain headposted and motionless in a custom-made fMRI chair while a RI-compatible
periscope presented a movie clip, a gray background (FOV 30 x 23, 60 Hz, eff. res. 530 400 fibers) or nothing. After the behavioral training was completed, the monkeys were scanned under the same conditions in a Bruker 4.7 T vertical MRI scanner with a custom-designed whole-head coil (single-shot GE-EPI, TR 1000 ms, TE 18 ms, 128 64 18 voxels, 1 1 2 mm).
Each run lasted 10 min (600 volumes). We collected 30 functional runs of resting state activity (without any visual stimulation) and 30 functional runs of stimulus driven activity (1 min of a natural movie presentation alternated with 1 min of gray background) for each monkey. All the volumes containing artifacts were pre-selected and excluded from the data analysis. For the visual stimulation condition, we selected the scans with strong visual activation based on a generalized linear model (GLM).
Functional connectivity data analysis (group-level ICA with 20 components) of the scans devoid of stimulation revealed resting-state networks consistent with previous reports in humans and monkeys (Mantini et al., 2013, J. Neurosci.). Furthermore, preliminary analysis of the scans with visual stimulation revealed components reflecting visually driven networks. Currently, we are employing the eigenvector centrality mapping (ECM), which is a parameter-free effective connectivity method (Lohmann et al., 2010, PLoS ONE) as well as models based dynamic causal modeling (DCM) (Friston et al., 2003, Neuroimage) to delineate differences across stimulation with different contexts and to characterize the physiological mechanisms behind the
transition of brain states.