hide
Free keywords:
-
Abstract:
Introduction
Spectral analysis in resting-state fMRI (rs-fMRI) studies has mainly focused on the 0.01 to 0.1 Hz frequency range1-6. Frequencies under 0.01 Hz are typically regarded as artifacts from scanner instabilities or physiological noise7, and are routinely excluded from the rs-fMRI analysis. Here, we show robust ultra-slow rs-fMRI signal fluctuations of high regularity in the brain of rats under anesthesia, which seem to be independent from breathing-derived motion8 or cardiovascular oscillations9,10 and possibly indicate a peculiar brain state in the animals.
Methods
12 to 15 minutes of rs-fMRI data were acquired from anesthetized rats (under isoflurane, a-chloralose, medetomidine or urethane) using a 3D-EPI sequence with the following parameters: TE, 12.5 ms; TR, 1s; matrix size, 48x48x32; resolution, 400x400x600 µm. All images were acquired with a 12 cm diameter 14.1 T/26 cm magnet interfaced to an Avance III console. Trans-receiver surface coils were used to acquire the whole brain fMRI. Animals were mechanically ventilated, and a low dose of the paralytic agent pancuronium was used to prevent motion artifacts. Frequency decomposition analysis and power estimation of the 0.005-0.012 Hz frequency band were performed to map the ulstra slow oscillations (USO) in the rat brain.
Results/Discussion
Here we show that frequencies below the classical 0.01 Hz limit can be detected, with high amplitude and rhythmic pattern, in the Blood-Oxygen Level Dependent (BOLD) fMRI signal of animals receiving anesthesia (Fig.1). When present, these slow waves occured predominantly in the hypothalamic area, as confirmed with bandpass power analysis (Fig.2). The features of the reported USO (brain region predominance and apparent absence of correlation with motion or physiological artifacts), suggest that these oscillations might have a neural origin instead of being derived from MR hardware noise. Importantly, infraslow oscillations in a similar range were detected by EEG in brain-injured patients, which appeared related to modulations in the cortical excitability and have been hypothesized to emerge from a deep brain source11.
Conclusions
Our observations suggest that frequency components in the ultra-slow range may contribute to brain function. Future work will aim to clarify the source of these oscillations with neuronal and astrocytic calcium imaging and the utilization of cholinergic modulators to study reversibility of these oscillations in the rat brain.
