Intraoperative Infrared Functional Imaging of Human Brain Alexander M. Gorbach, PhD, 1,2 John Heiss, MD, 1 Conrad Kufta, MD, 1 Susumo Sato, MD, 3 Paul Fedio, PhD, 4 William A. Kammerer, MD, 5 Jeffrey Solomon, MS, 6 and Edward H. Oldfield, MD 1 We hypothesized that it would be possible to detect the distribution of cortical activation by using a sensitive, rapid, high-resolution infrared imaging technique to monitor changes in local cerebral blood flow induced by changes in focal cortical metabolism. In a prospective study, we recorded in 21 patients the emission of infrared radiation from the exposed human cerebral cortex at baseline, during language and motor tasks, and during stimulation of the contralateral median nerve using an infrared camera (sensitivity 0.02°C). The language and sensorimotor cortex was identified by standard mapping methods (cortical stimulation, median nerve somatosensory-evoked potential, functional magnetic resonance imaging), which were compared with infrared functional localization. The temperature gradients measured during surgery are dominated by changes in local cerebral blood flow associated with evoked functional activation. The distribution of the evoked temperature changes overlaps with, but extends beyond, functional regions identified by standard mapping techniques. The distribution observed via infrared mapping is consistent with distributed and complex functional representation of the cerebral cortex, rather than the traditional concept of discrete functional loci demon- strated by brief cortical stimulation during surgery and by noninvasive functional imaging techniques. By providing information on the spatial and temporal patterns of sensory-motor and language representation, infrared imaging may prove to be a useful approach to study brain function. Ann Neurol 2003;54:297–309 Although usually presented as a measure of neuronal activity, functional imaging techniques, including positron emission tomography, functional magnetic resonance imaging (fMRI), and optical intrinsic imag- ing, measure either the rate of consumption of an en- ergy substrate (glucose, oxygen) or neurophysiological responses to such consumption (rate of local cerebral blood flow [CBF]). Cortical regions of increased func- tional activity also are identified by energy released in the form of electrogenesis (action potentials) or intrin- sic production of heat. Correlation of the timing and localization of heat production of brain structures to functional activation has been previously examined. 1 Thermocouples have been used to measure stimulation-induced temperature shifts during presen- tation of visual, somesthetic, and auditory stimuli. 2–4 However, thermocouples compress adjacent vessels, modifying CBF-dependent local heat dissipation, and the studies are inconclusive for location of temperature foci, its spatial heterogeneity, and location of function. Gorbach mapped stimulation-induced temperature re- sponses in the brain using infrared imaging through the animal scalp and correlated stimulation of rat whis- kers with localization of the cortical temperature re- sponses. 5,6 However, these experiments did not deter- mine if the detected shifts in brain temperature were a result of changes in neuronal metabolism or cerebral blood flow. We hypothesized that evaporation from the exposed brain would increase the rate of heat dissipation, en- hancing discrimination of local thermal gradients and the capacity to localize cortical activity by neuronal heat production or by the increased capillary blood flow that is tightly linked regionally and temporally to increased neural firing. 7,8 Success would allow precise imaging of cortical functional activation in time and space. From the 1 Surgical Neurology Branch, National Institute of Neu- rological Disorders and Stroke; 2 Department of Radiology, Warren Grant Magnuson Clinical Center; 3 Office of Clinical Director, Na- tional Institute of Neurological Disorders and Stroke, National In- stitutes of Health, Bethesda, MD; 4 Department of Psychology, George Mason University, Fairfax, VA; 5 Department of Anesthesi- ology, Warren Grant Magnuson Clinical Center, National Institutes of Health, Bethesda, MD; and 6 Sensor Systems Inc., Sterling, VA. Received Nov 25, 2002, and in revised form Feb 25 and Apr 11, 2003. Accepted for publication Apr 11, 2003. Address correspondence to Dr Oldfield, National Institutes of Health, NINDS, Surgical Neurology Branch, Building 10, Room 5D37, Bethesda, MD 20892. E-mail: eo10d@nih.gov This article is a US Government work and, as such, is in the public do- main in the United States of America. Published by Wiley-Liss, Inc., through Wiley Subscription Services 297