Purpose

Rats have long been utilized for fMRI experiments ranging from basic neurophysiological studies to translational imaging of animal models. Anesthesia is widely used to prevent motion and ideally provide stable physiological conditions. Imaging unanesthetized rats avoids anesthesia effects on neural activity and the vasculature, but introduces challenges due to potential issues arising from rigid restriction [1]. So far only a few groups have reported awake rat fMRI studies [2,3]. What is the major limitation for its broad application in the research community? To answer this critical question, we adopted the restriction strategy suggested in [3], a common paradigm for preparing animals for awake fMRI. We evaluated the frequency of struggling and the condition of the rat during daily restriction training and conducted a preliminary comparison of fMRI resting-state data between the waking state and deep anesthesia.

Methods

Eleven Sprague-Dawley rats (~250g, male) were included in this study; 4 were used as controls and 7 underwent acclimation training (restriction/noise) for 7-9 consecutive days with a restraining system (Animal Imaging Research, MA), according to the procedure in [2]. For fMRI studies, the trained animals underwent ~8-min resting state scans during 2% isoflurane (ISO) administration and in the awake state (9.4T Bruker MRI, GE-EPI, TR=500 ms, TE=15 ms, 5 2-mm slices). The data were prepared in SPM (slice timing correction, motion correction and spatial smoothing with 2.5X voxel size FWHM), and bandpass filtered to 0.01-0.15 Hz for functional connectivity analysis. Correlation maps (with or without global signal regression) and power spectra were calculated for seeds in somatosensory cortex and compared across conditions.

Results

During the daily acclimation, weight gain slowed (2-8g per day) relative to control rats (>12g per day). In 6 of 7 rats, the skin of areas near the ear bars appeared slightly damaged and there were scars at the end of training. Struggling became infrequent after a couple days’ acclimation, with some rats (4 of 7) able to remain still more than consecutive 15min without visible struggling by the end of the training. These rats were considered ready for the resting-state fMRI scan. However, no animal were observed that had no struggling movements during a 30-60 min restriction at the end of acclimation training. One or more 5-10min resting state fMRI awake data sets without struggling movement were obtained from 3 of the 7 trained animals. The data acquired during this stable period had minimal head motion, less than ¼ voxel size (0.4*0.4*2mm³), similar to the data acquired under deep isoflurane (fig. 1). Raw time courses are shown in Fig. 2 and power spectra and functional connectivity maps are shown in Fig. 3.

Discussion/ Conclusion

Struggling movements were obvious in the initial days of acclimation, which may increase the risk of pain during restriction, especially near areas where the ear-bars make contact. Ear bars are key for high quality immobilization, but need a proper force to avoid skin damage during struggles. We tried replacing the small-tipped ear bar with an ear pad and found no skin damage but the head position could be changed after a struggling movement. Rats exhibited a reduction in struggling frequency following acclimation. The comparison of data from anesthetized and awake animals shows clear differences in functional connectivity and the effect of global signal regression. Interestingly, global signal regression increased the resemblance between the data acquired under isoflurane and the data acquired in awake animals, similar to previous findings that global signal regression minimized the effects of varying doses of anesthesia [3]. Studies in humans have also found that the global signal is tied to alertness and that caffeine can have similar effects to GSR [4]. Anesthesia is convenient and provides a stable state for animal experiments, but alters brain activity. Awake rat fMRI avoids the confounding effects of anesthesia on neural activity but is time-consuming, possibly stressful and/or painful, and has highly variable results across rats. Until a pain-free restriction method that introduces minimal stress is developed, the assumption that awake rat fMRI represents a “normal” brain state is not well supported. Further efforts are still needed to improve the methods of restriction and acclimation. In the meanwhile, global signal regression should be further investigated as a tool to make data acquired in anesthetized animals more similar to data acquired from awake rats.

Acknowledgements

NIH, R01 NS078095