Preclinical functional magnetic resonance imaging (fMRI) is a powerful brain research tool, as it can be used in more controlled and invasive experiments compared to human studies. In most animal fMRI study settings, however, anesthesia is a major confounding factor, because the anesthesia directly affects brain function and responsiveness to stimuli.¹ Recent evidence suggests that parameters obtained with resting-state fMRI (rs-fMRI) are coupled with anesthetic depth.² Therefore, we investigated whether rs-fMRI parameters, such as functional connectivity (FC), also have a relation to blood oxygen level dependent (BOLD) responses to different stimuli. A simple rs-fMRI protocol was implemented in a pharmacologic fMRI (phMRI) study evaluating brain hemodynamic responsiveness and anesthetic state under five anesthetics. The temporal change in FC was also evaluated at 1-h interval during the experiments.The animal procedures were approved by the National Animal Experiment Board. Male Wistar rats (347±36 g) were used in the fMRI experiments with five anesthetics: α-chloralose (n=8, 60 mg/kg + 30 mg/kg/h i.v.), isoflurane (n=8, 1.3%), medetomidine (n=7, 0.1 mg/kg/h i.v.), thiobutabarbital (n=10, 140 mg/kg i.p.), and urethane (n=8, 1.25 g/kg i.p.). All rats were anesthetized with isoflurane during femoral artery and vein cannulation. A tracheal tube was inserted for ventilation. The fMRI data were acquired with 7T Bruker Pharmascan with single-shot spin-echo echo planar imaging (TR 2 s, TE 45 ms, FOV 2.5x2.5 cm2, 64x64 matrix, and 9x1.5 mm slices). The fMRI session included two rs-fMRI acquisitions (300 volumes) with 1-h interval and an acute nicotine (88 µg/kg i.v.) challenge phMRI scan (700 volumes) between rs-fMRI measurements. Additionally, forepaw stimulation was performed for two TBB-anesthetized rats. Arterial blood samples were obtained and analyzed twice during imaging. The MRI data were converted to NIfTI (http://aedes.uef.fi), slice-timing corrected, motion-corrected, spatially smoothed, and co-registered using SPM8 (www.fil.ion.ucl.ac.uk/spm) and Matlab (Version 2011a, The Mathworks Inc., Natick, MA, USA). The analyses were performed using in-house Matlab code, Aedes, and SPM8.Physiologic parameters were in normal range. The highest nicotine-induced BOLD responses were observed in the urethane group. Rats under medetomidine and isoflurane anesthesia also showed robust responses, while under α-chloralose anesthesia BOLD responses evoked by nicotine were weak. Within the thiobutabarbital group we observed two different response patterns (Figure 1), and correspondingly the animals were stratified into two subgroups: Subgroup 1 (TBB-SG1) and Subgroup 2 (TBB-SG2). The coherence analysis of the resting-state acquisitions revealed similar grouping patterns among the thiobutabarbital rats (Figure 2). Spearman’s rank correlation coefficients (ρ) between baseline FC and nicotine-induced BOLD revealed significant positive correlations in the α-chloralose, isoflurane, medetomidine, and thiobutabarbital groups (Table 1). The temporal changes in global FC are shown in Figure 3. In α-chloralose, TBB-SG2, and urethane groups the FC values significantly increased during the 1-h observation period. The permanence of the coupling between FC and hemodynamic responsiveness was investigated with somatosensory stimulation experiments (Figure 4). Rat with initially weak coherence values and negative nicotine response also had a weak BOLD response to forepaw stimulation. However, at 2^nd time point, the coherence values were increased and a robust BOLD response to forepaw stimulation was detected.The results show that FC measured during baseline period was able to predict the magnitude of BOLD response to nicotine stimulation, especially under thiobutabarbital anesthesia. More importantly, the resting-state data was highly valuable during the data interpretation. For example, the measured standard physiologic parameters were not able to distinguish the subjects with different responsiveness to nicotine. Accurate measurement of the anesthetic depth would require more direct methods, such as electrophysiologic recordings, which are technically demanding in the magnet bore. In all groups in which a significant temporal change in FC was observed, the anesthetic was administered as a bolus. With the continuously administered anesthetics, isoflurane and medetomidine, significant differences were not detected between the two time points. This observation might suggest that FC dynamically changes according to the pharmacodynamics of the anesthetic, and an rs-fMRI protocol is able to detect these changes.Our findings indicate that rs-fMRI is capable of evaluating brain hemodynamic responsiveness to pharmacological and sensory stimulation and temporal progression of anesthesia. Therefore, FC analysis of baseline data is a highly valuable and readily implementable tool for controlling preclinical fMRI experiments.