fMRI Evaluation of Somatotopic Representation in Human Primary Motor Cortex M. Lotze,* , † M. Erb,* H. Flor,‡ E. Huelsmann,* B. Godde,† and W. Grodd* *Section for Experimental Magnetic Resonance of the Central Nervous System, Department of Neuroradiology, and †Institute for Medical Psychology and Behavioral Biology, University of Tu ¨ bingen, D-72074 Tu ¨ bingen, Germany; and ‡Department of Psychology, Humboldt-University, Berlin, Germany Received July 26, 1999 We used fMRI to map foot, elbow, fist, thumb, index finger, and lip movements in 30 healthy subjects. For each movement type confidence intervals of represen- tational sites in the primary motor cortex (M1) were evaluated. In order to improve the precision of their anatomical localization and to optimize the mapping of cortical activation sites, we used both the assess- ment of locations in the conventional 3D system and a 2D projection method. In addition to the computation of activation maxima of activation clusters within the precentral gyrus, centers of gravity were determined. Both methods showed a high overlap of their repre- sentational confidence intervals. The 2D-projection method revealed statistically significant distinct in- tralimb locations, e.g., elbow versus index finger move- ments and index finger versus thumb movements. In- creased degree of complexity of finger movements resulted in a spread of the somatotopic location to- ward the arm representation. The 2D-projection meth- od-based fMRI evaluation of limb movements showed high precision and was able to reveal differences in intralimb movement comparisons. fMRI activation re- vealed a clear somatotopic order of movement repre- sentation in M1 and also reflected different degrees of complexity of movement. © 2000 Academic Press INTRODUCTION Somatotopy in the Primary Motor Cortex Jackson (1873/1958) postulated distinct cortical ac- tivation centers for different types of movements after having observed isolated movements of different body parts during epileptic convulsions. He further hypoth- esized that the size of the cortical area which repre- sents movements is not proportional to the size and strength of the muscle group involved in the movement but to the degree of differentiation and specialization of the type of movement. These ideas of Jackson were directly tested by electrical stimulation of the cortex in the second part of the 19th century. Fritsch and Hitzig (1870) used galvanic stimulation of a dog’s brain to show that specific movements (forelimb extension) could be elicited by stimulation of a particular cortical location. Ferrier (1875) extended these results to mon- keys and discovered a precise somatotopic map of movement representations. The differentiation of pre- and postcentral gyri, the former as motor—the latter as sensory— cortex, was first observed by Gruenbaum and Sherrington (in Leyton and Sherrington, 1917). Foerster (1936) as well as Penfield and Boldrey (1937) described a somatotopic order of movements in the human precentral gyrus comparable to the somato- topic representation of somatosensory stimuli in the postcentral gyrus. A review of somatotopic representa- tions based on cortical microstimulation is given in Phillips and Porter (1977). More precise electrical stimulation techniques in monkeys (e.g., Kwan et al., 1978; Nudo et al., 1992) suggest that the concept of a somatotopic organization of the primary motor cortex (M1) is probably adequate only in a gross approximation. The representation ar- eas of finger and hand movements in M1 show strong overlap, indicating that movements of different repre- sentational sites share at least some neural substrates. This overlap may be due to a lack of isolation of differ- ent movements. For instance, Schieber and Hibbard (1993) reported that monkeys trained to execute finger movements show comovements of nonparticipating fin- gers. Furthermore, the functional overlap may also be due to stabilization of distal movements by the more proximally located muscle groups. Representational areas of different limbs and body parts which do not share functional coactivation, as for instance during proximal stabilization, rarely overlap (Huntley and Jones, 1991). Schieber and Hibbard (1993) argued that patches of neuronal groups may enable different move- ments of one limb embedded in a functional integrity. Each finger movement appears to be specified by a neuronal population distributed throughout the M1 hand area. Lesion studies also suggest that neuronal groups in M1 are represented in functional integrities NeuroImage 11, 473– 481 (2000) doi:10.1006/nimg.2000.0556, available online at http://www.idealibrary.com on 473 1053-8119/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.