IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 44, NO. 11, NOVEMBER 1997 1107 Automated Feedback Control of Body Temperature for Small Animal Studies with MR Microscopy Hui H. Qiu, Gary P. Cofer, Laurence W. Hedlund*, and G. Allan Johnson, Associate Member, IEEE Abstract—A temperature control system consisting of a ther- mistor, signal processor, and computer algorithm was developed for magnetic resonance (MR) microscopy of small live animals. With control of body temperature within 0.2 C of the set point, heart rate is stabilized and, in turn, repetition time (TR) during cardiac-gated studies is less variable. Thus, image quality and resolution are improved. Index Terms— Body temperature control, computer control, feedback control, magnetic resonance microscopy, rodents. I. INTRODUCTION M OTION is the single largest barrier to increased spatial resolution for many in-vivo magnetic resonance (MR) microscopy studies. Major sources of biological motion are breathing and cardiac activity. A number of techniques for reducing motion artifacts have been used, such as respiratory and cardiac gating [1]–[4]. For extended periods of imaging, slower physiologic changes also become critical because, during cardiac-gated imaging, repetition time (TR) may vary during data acquisition and this can result in signal intensity variation and image artifacts. In addition to the relatively slow changes in imaging time from drifts in temperature and heart rate, TR may change by multiples of the cardiac pulse interval and this can lead to severe image degradation. To detect and compensate for slow drifts of body temperature and heart rate as well as maintain constant TR, it is necessary to carefully monitor physiologic parameters and attempt to maintain relative physiologic constancy. Physiologic stability also impacts on interstudy variability as well as intrastudy variability. Although physiologic parameters such as body temperature and heart rate can be maintained manually by changing heating and anesthesia levels, this is labor-intensive and not highly reliable. Several investigators have used automated control systems to heat or cool local tissue or nerve preparations [5], [6]. However, such previously published methods are not compatible with the severe constraints of magnetic resonance imaging (MRI), i.e., high magnetic and radio frequency fields and restricted access. To achieve greater physiologic stability Manuscript received July 16, 1996; revised March 26, 1997. This work was supported by the National Institutes of Health under NCRR Grant P41 RR05959. Asterisk indicates corresponding author. H. H. Qiu, G. P. Cofer, and G. A. Johnson are with the Center for In- Vivo Microscopy, Department of Radiology, Duke University Medical Center, Durham, NC 27710 USA. *L. W. Hedlund is with the Center for In-Vivo Microscopy, Department of Radiology, Duke University Medical Center, Box 3808, Durham, NC 27710 USA (e-mail: lwh@orion.mc.duke.edu). Publisher Item Identifier S 0018-9294(97)07412-0. (a) (b) Fig. 1. Photograph of the controller and heater/blower. (a) Top view of the controller; front is at bottom of picture. (b) Controller, heater/blower, and duct. and ease of animal support, we have designed and built, from commercially available components, an MR-compatible computer feedback control system for maintaining animal body temperature to provide physiologic stability of small animals during extended MRI sessions. The characteristics of this system were evaluated during in-vivo studies in rats, guinea pigs, ferrets, and mice. II. METHODS Temperature Control System Overview: Two identical con- trol systems were installed on two small-bore MRI sys- 0018–9294/97$10.00  1997 IEEE