Fig 1.
Experimental design for fMRI during VNS.
Each rat was stimulated at the left cervical vagus through a cuff electrode implanted in an acute surgery. Biphasic current pulses were delivered during a 10s “ON” period alternating with a 50s “OFF” period for 10 cycles. With this block design, the rat was scanned for fMRI with a repetition time of 1s.
Fig 2.
VNS-evoked responses varied across regions.
(A) shows the response time series averaged within each of the three regions of interest: the retrosplenial cortex (RSC) (blue), the brainstem (green), and the dorsal caudate putamen (Cpu) (red). (B) shows the highly different activation maps based on the response models derived with the HRF, for which the peak latency was assumed to be 3s, 6s, or 9s. The color shows the group average of the z-transformed correlation between the voxel time series and the modeled response. The maps were thresholded with p<0.05 (one-sample t-test, uncorrected).
Fig 3.
VNS evoked widespread and complex responses in the brain.
(A) VNS-evoked responses for different brain networks derived with ICA. The ICA-defined networks are labeled as: amygdala (Amy), caudate putamen (Cpu), hippocampus, (Hipp), cingulate cortex (Cing), prelimbic cortex (PrL), infralimbic cortex (IL), brain stem, hypothalamus (HTh), thalamus (Tha), superior colliculus (SC), cerebellum (Cb), primary and secondary motor cortex (M1, M2), and primary and secondary somatosensory cortex (S1, S2). For each network, the time points at which the responses were significant are shown in red. (B) The VNS-activated voxels cover 76.03% of the brain volume. The color represents the standard deviation of the voxel-wise response averaged across repetitions of VNS. The locations with the greatest responses are highlighted with arrows. Data relevant to the VNS-evoked network responses are available in the online Supplementary Information.
Fig 4.
VNS altered the functional connectivity among functional networks.
(A) shows the correlations between independent components. The left shows the correlation matrix during the resting state (or the “control” condition). The right shows the correlation matrix during VNS (or the “VNS” condition). Smaller squares highlight the networks (or ICs) that were clustered into groups (based on k-means clustering). (B) shows the IC-IC functional connectivity that was significantly different between the VNS and control conditions (t-test, P<0.005). Red lines represent increases in functional connectivity, and green lines represent decreases in functional connectivity. The thickness of the lines represents the (VNS minus control) change in correlation. The brain maps show the spatial patterns of individual ICs. Corresponding to the squares in (A), the arc lines illustrate how the ICs were clusters into groups, for the VNS condition (inner circle) and the control condition (outer circle).