Resting state network (RSN), the low frequency spontaneous fluctuation of the BOLD signal [1], has been of interest in neuroimaging for many purposes. RSNs have been identified, which are associated with sensory, motor or other cognitive functions, although they are not directly perceptible for us. Usually, the subjects are instructed not to do anything and keep rest, however, the confirmation of keeping rest, i.e. the behavioral data during the ‘rest performance’, mostly depended on the self-reports by the subjects. In this study, we applied monitoring of autonomic responses, the pupil diameter during RS measurement, to evaluate the RS of the subjects. This method directly detects the status of an eye representing the consciousness level. The relationship between pupil diameter or eye openness and the RSN activity was investigated.Thirty-one young subjects (Age 20 – 36, 18 males) who gave written informed consent participated in this study. The subjects performed RS fMRI measurements on a 3T MRI scanner. They were instructed to fix their eyes to a cross mark attached on the internal wall of the magnet’s bore and do nothing else during the session. A RS-fMRI session continued for 7 minutes by using a T2* weighted gradient recalled echo EPI sequence with the following parameters: TR 3000 ms, TE 30 ms, 39 axial slices, 3 mm thick with distance factor of 25%, FOV 19.2 cm. The diameter of the left pupil of a subject was recorded by using a CCD camera (CAM- LED-IR, MRC) with near-infrared light source. The recorded video images (640x480, 30Hz) were processed and the time courses of pupil diameters were extracted by using a software tool designed for pupil dynamic analysis (PiPS, AIST). The functional images were pre-processed and statistically analyzed by SPM12 (UCL) and the RSN activity was extracted by GIFT4.0 for a group ICA analysis (UNM). Shirer’s template [2] was used to assign the components. The final contrast maps were obtained by a 2x2 factorial design of the two factors, pupil diameter and eye-openness. The time courses of the pupil diameter were classified into two groups, constant type (17 subjects, Fig.1A) and decay type (15, Fig1B) depending on the gradient of regression line. The maximum decay at the end of measurements was 40% compared with the diameter at the beginning. No subject presented increasing of pupil diameter. The eye openness was evaluated by the number of 0 value, which mostly represented errors to detect the pupil edge due to eye closing including blinking or miss-registration of the pupil edge. Twenty-two subjects were classified into stable eye-open and the other 10 into frequent close by eye-openness ratio (EOR). Twenty-five components were extracted and the 14 components defined by Shirer’s template were determined by multiple regression. With a higher threshold (FWE, p<0.05), no significant difference of the main effect, interaction and positive effects was detected among the 4 classes in all the 25 components. With compromised threshold (p<0.01, uncorrected), differential contrast was obtained at posterior cingulate of the ventral default mode network (DMN). The differential peak coordinate of [constant – decay] was [-2 -50 48] (p < 0.001, T = 3.97) (Fig.2A) in eye open group and that of [decay – constant] was [2 -56 56] (p < 0.001, T = 3.49) in frequent eye-close group (Fig.2B). These two clusters did not overlap each other, although they were very closely located.It was confirmed that the pupil diameters of 47% of the subjects who could keep their eyes open during the RS-fMRI session attenuated during the first 3 minutes, suggesting that these subjects were tensioned in the beginning and then relaxed. No significant effect of pupil diameter stability or eye openness on RNS activation with conservative threshold indicated that the RSNs are robust and not strongly affected by tension or eye closing for short time. Since DMN, especially the posterior cingulate gyrus is associated with consciousness [3], less activation of this area in decay type may suggest that the subjects of constant type kept higher tension until the end of the session, rather than they are relaxed from the beginning. Considering that the extent of subjects’ tension may change in the first 3 minus, measurements longer than 5 minutes or preliminary relaxation time before starting scanning may represent RS with less intension to keep eyes open or keep RS, although such ‘resting task’ does not give significant biases to the RSNs except to the DMN.