900 nature neuroscience • volume 5 no 9 • september 2002 articles It is thought that human tactile sensation is mediated exclusively by large myelinated afferents, whereas pain and temperature sensations are mediated by small myelinated and unmyelinated afferents 1 . Nev- ertheless, it is known that the skin of humans and other mammals is supplied with a system of slow-conducting unmyelinated affer- ents that respond vigorously to innocuous skin deformation 2–8 . CT afferents have been found in peripheral nerves from several skin areas but are lacking in glabrous (hairless) skin 9–12 . Their functional properties differ from those of myelinated afferents: they respond particularly well to slow stroking of the skin, but poorly to rapid skin deformation, including vibration 8,10,13 . Animal studies show that they project to the outer layers of the spinal dorsal horn 14–17 . The functional role of the CT afferents is not known 18 but their response properties and slow conduction velocity suggest that they are not important for cognitive aspects of tactile stim- ulation. Rather, it seems likely that they are involved in limbic functions, particularly emotional aspects of tactile stimula- tion 3,5,19 . Direct evidence is difficult to obtain, however, as tac- tile stimuli always activate the myelinated system in parallel. An opening has now been provided by a unique patient, G.L. (see ref. 20 and Methods), with selective loss of large-diameter myeli- nated afferents. We report that this patient was able to detect tac- tile stimuli that were designed to activate CT afferents, suggesting that these afferents may serve tactile sensation. Moreover, fMRI analysis indicated that CT afferents have strong projections to the insular region, consistent with a role in emotional touch. RESULTS In a two-alternative, forced-choice (2AFC) test, the patient G.L. could detect light touch (stroking with a soft brush) applied to the Unmyelinated tactile afferents signal touch and project to insular cortex H. Olausson 1 , Y. Lamarre 2 , H. Backlund 1,3 , C. Morin 4 , B.G. Wallin 1 , G. Starck 5 , S. Ekholm 6 , I. Strigo 4 , K. Worsley 7 , Å.B. Vallbo 3 and M.C. Bushnell 4 1 Department of Clinical Neurophysiology, Sahlgrenska University Hospital, S-413 45 Göteborg, Sweden 2 Department of Physiology, University of Montreal, Montreal, Quebec H3C 3J7, Canada 3 Department of Physiology, Göteborg University, S-413 45 Göteborg, Sweden 4 Department of Anesthesia, McGill University, Montreal, Quebec H3G IY6, Canada 5 Department of Radiation Physics, Sahlgrenska University Hospital, S-413 45 Göteborg, Sweden 6 Department of Radiology, Sahlgrenska University Hospital, S-413 45 Göteborg, Sweden 7 Department of Mathematics and Statistics, McGill University, Montreal, Quebec H3G IY6, Canada Correspondence should be addressed to H.O. (olausson@physiol.gu.se) Published online: 29 July 2002, doi:10.1038/nn896 There is dual tactile innervation of the human hairy skin: in addition to fast-conducting myelinated afferent fibers, there is a system of slow-conducting unmyelinated (C) afferents that respond to light touch. In a unique patient lacking large myelinated afferents, we found that activation of C tactile (CT) afferents produced a faint sensation of pleasant touch. Functional magnetic resonance imaging (fMRI) analysis during CT stimulation showed activation of the insular region, but not of somatosen- sory areas S1 and S2. These findings identify CT as a system for limbic touch that may underlie emo- tional, hormonal and affiliative responses to caress-like, skin-to-skin contact between individuals. forearm and the dorsum of the hand, but she could not detect this stimulus on the glabrous skin (palm) of the hand (Fig. 1a). When G.L. was specifically questioned after the test session, she reported that, if she really concentrated, she was able to perceive a faint and diffuse touch sensation, which she described as “a pressure” applied to the arm or dorsum of the hand. Without knowing what kind of stimulus we delivered, she reported that the percept was clearly pleas- ant with no sensation of pain, temperature, itch or tickle. She per- ceived the brush-evoked sensation on the forearm as significantly weaker than did normal subjects, but pleasant to a similar degree as did normal subjects (Fig. 1f). When an intact skin area (forehead) was stimulated, she perceived the brushing as equally intense and pleasant as did normals (Fig. 1f). G.L. could not tell the direction of a probe moving for a long distance (100 mm) along the forearm (48% correct in 2AFC procedure), whereas normal subjects readi- ly perceive the direction of 10 mm movements 21 . In contrast to her detection of the moving brush, G.L. could not detect a local vibra- tion applied to the forearm, hand dorsum or palm (Fig. 1b). G.L. and normals had similar thresholds for detection of warmth and cold pain in all tested areas (Fig. 1c and d), suggest- ing well-functioning unmyelinated fiber systems. But G.L. had higher thresholds than normals for cool detection in the dorsum of the hand and forearm (Fig. 1e), suggesting a partial distur- bance of thin myelinated fiber systems. We used fMRI to explore brain response to CT stimulation. In normal subjects, stroking with a soft brush on the right fore- arm activated contralateral primary (S1) and bilateral secondary (S2) somatosensory cortices, and contralateral insular cortex (IC) (Table 1 and Fig. 2). There was one exception to the pattern: one normal subject (I.L.) showed no significant activation in S1. In © 2002 Nature Publishing Group http://www.nature.com/natureneuroscience