Functional topography of early periventricular brain lesions in relation to cytoarchitectonic probabilistic maps

Transcription

1 Available online at Brain & Language 106 (2008) Functional topography of early periventricular brain lesions in relation to cytoarchitectonic probabilistic maps Martin Staudt a,b, *, Luca F. Ticini c, Wolfgang Grodd b, Ingeborg Krägeloh-Mann a, Hans-Otto Karnath c a Department of Pediatric Neurology and Developmental Medicine, University Children s Hospital, Hoppe-Seyler-Str. 1, D Tübingen, Germany b Section Experimental MR of the CNS, Department of Neuroradiology, Radiological Clinic, University of Tübingen, Tübingen, Germany c Section Neuropsychology, Center for Neurology, Hertie-Institute for Clinical Brain Research, University of Tübingen, Tübingen, Germany Accepted 13 January 2008 Available online 13 February 2008 Abstract Early periventricular brain lesions can not only cause cerebral palsy, but can also induce a reorganization of language. Here, we asked whether these different functional consequences can be attributed to topographically distinct portions of the periventricular white matter damage. Eight patients with pre- and perinatally acquired left-sided periventricular brain lesions underwent focal transcranial magnetic stimulation to assess the integrity of cortico-spinal hand motor projections, and functional MRI to determine the hemispheric organization of language production. MRI lesion-symptom mapping revealed that two distinct portions of the periventricular lesions were critically involved in the disruption of cortico-spinal hand motor projections on the one hand and in the induction of language reorganization into the contra-lesional right hemisphere on the other hand. Both regions are located in a position compatible with the course of cortico-spinal/cortico-nuclear projections of the primary motor cortex in the periventricular white matter, as determined by the stereotaxic probabilistic cytoarchitectonic atlas developed by the Jülich group. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Early brain lesions; Reorganization; Cerebral palsy; Functional MRI; Transcranial magnetic stimulation; Cortico-spinal tract 1. Introduction Periventricular brain lesions are among the most frequent causes for cerebral palsy (Ashwal et al, 2004). This type of lesion typically occurs during the early 3 rd trimester of pregnancy, either as intrauterine insults or as complications of premature birth (Krägeloh-Mann, 2004; Staudt et al, 2004). The motor dysfunction of patients with such lesions is commonly believed to result from structural damage to cortico-spinal projections in the periventricular white matter (Banker & Larroche, 1962). Indeed, several transcranial magnetic stimulation (TMS) studies could show that in patients with large unilateral periventricular lesions, TMS of the affected hemisphere did not elicit * Corresponding author. Fax: address: martin.staudt@med.uni-tuebingen.de (M. Staudt). motor responses in target muscles of the (contralateral) paretic hand, indicating that the lesion had disrupted the normal crossed cortico-spinal projections from the affected hemisphere (Carr, Harrison, Evans, & Stephens, 1993; Maegaki et al, 1997; Nezu, Kimura, Takeshita, & Tanaka, 1999; Staudt et al, 2002a). Periventricular lesions, however, do not only affect the motor system, but can in the case of left-sided lesions also induce a reorganization of productive language functions into the contralesional, right hemisphere (Staudt et al, 2001, 2002b). Right-hemispheric language organization induced by early left-sided brain lesions typically achieves normal language functions (Table 1; Muter, Taylor, & Vargha-Khadem, 1997; Staudt et al., 2002b), but can compromise right-hemispheric functions such as visuo-spatial abilities (Lidzba, Staudt, Wilke, & Krägeloh-Mann, 2006). To observe right-hemispheric organization in our X/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi: /j.bandl

2 178 M. Staudt et al. / Brain & Language 106 (2008) Table 1 Patients demographic and clinical data, pre- and perinatal history, results of the TMS investigation (preserved vs. disrupted cortico-spinal projections from the affected hemisphere to the paretic hand), of the fmri investigation [number of activated voxels in left (L) and right (R) hemisphere; SPM99, p < 0.05 corrected for multiple comparisons at voxel level], and verbal IQ scores (using the German adaptation of the Wechsler scale; Tewes, 1994) Verbal IQ Number of activated voxels Language Laterality (fmri) Cortico-spinal tract integrity (TMS) Handedness Hand motor dysfunction Pre- and perinatal complications Birth weight (g) Pt Age (y) Sex Prematurity (GA) L R 1 23 m 3800 Cerclage at 33 w GA L 1 Preserved L > R m 3650 L 1 Preserved L > R f 3040 L 1 Preserved R > L m 33 w 2380 L 2 Preserved R > L f 3340 L 2 Disrupted L > R f 3400 L 2 n/a R > L n/a 6 18 m 2300 Abnormal CTG, CS L 3 Disrupted R > L f 36 w 2500 L 2 Disrupted R > L Hand motor dysfunction was graded with the sequential finger opposition task as reported previously (Staudt et al, 2002a), with 1 = normal performance; 2 = slow and/or incomplete performance; 3 = inability to perform any independent finger movements, but with a preserved grasp function. The patients are numbered as in our previous publication (Staudt et al, 2001), but sorted according to the TMS and fmri findings. GA, gestational age; CTG, cardiotocogram; CS, Caesarean section; w, weeks; n/a, not available. patients with periventricular lesions was unexpected insofar as these lesions typically leave cortical language zones at least macroscopically intact. As a possible explanation for this phenomenon, we suggested that structural damage to facial motor projections of the left hemisphere, with their well-known relevance for articulation (Wildgruber, Ackermann, Klose, Kardatzki, & Grodd, 1996), might induce this reorganization of language. In the current study, we addressed this issue by asking whether specific, topographically distinct locations of periventricular lesions are critical for (a) the disruption of hand motor projections of the cortico-spinal tract and (b) the induction of language reorganization. We did this by taking advantage of previously collected and published data on this well-described sample of patients (Staudt et al., 2001, 2002a). 2. Materials and methods 2.1. Subjects Eight patients (4 women; age range, years) with congenital right hemiparesis due to left-sided periventricular brain lesions were included (Table 1). Informed written consent and approval from the local ethics committee were obtained. In a previous study (Staudt et al., 2002a), seven of these patients had undergone focal TMS to determine whether the affected hemisphere still possessed crossed corticospinal projections to the paretic hand. Four patients showed preserved crossed projections to the paretic hand, three did not. No TMS data were available for patient #4 (Table 1); the subject thus was excluded from the analysis on cortico-spinal projections (see below). In a further study (Staudt et al., 2001), all but one patient (#8) from the current sample were examined with respect to right- and left-hemisphere activation during speech by functional MRI during silent word generation: The subjects were instructed to silently generate word chains, the first word starting with a letter given by the examiner, the next word starting with the last letter of the previous word, and so on. The procedure, which was applied in an identical manner in patient #8, is described in detail in Staudt et al. (2001). The lateralization of language production in this group of patients was found to be highly variable (Table 1) in contrast to healthy righthanded controls, who showed a consistent and strong leftward asymmetry of activation. Due to the overall sample size, we dichotomized the patient group into those presenting a higher number of fmri activated voxels in the left versus the right hemisphere (L > R; n = 3), and those with a higher number in the right versus the left hemisphere (R > L; n =5;Table 1) Lesion analysis The brain lesions of the patients were demonstrated by structural MRI (Fig. 1). Anatomical 3D data sets were

3 M. Staudt et al. / Brain & Language 106 (2008) obtained on a 1.5 Tesla Siemens Vision system (MPRAGE, 128 contiguous sagittal slices, TR [repetition time] = 9.7 ms, TI [inversion time] = 300 ms, TE [echo time] = 4 ms, voxel size mm 3.) Mapping of lesions was carried out by one experimenter without knowledge of test results and clinical features of the patients. The boundary of the lesion (including periventricular white matter loss and ventricular enlargement) was delineated manually on every single transversal slice of the individual MRI using MRIcro software (Rorden & Brett, 2000) ( Periventricular white matter loss was determined by delineating the enlarged portion of the left ventricle through direct comparison with the unaffected right ventricle shape on the same transversal slice. Both the scan and lesion shape were then transformed into stereotaxic space using the spatial normalisation algorithm provided by SPM2 ( using default settings. For determination of the transformation parameters, cost-function masking was employed (Brett, Leff, Rorden, & Ashburner, 2001). None of the patients showed a midline shift, neither before nor after normalisation (Fig. 1). To investigate whether topographically different portions of a periventricular lesion are responsible for a disruption of cortico-spinal hand motor projections and for the reorganization of language into the contra-lesional right hemisphere, we classified the patients into different groups and performed group contrasts using lesion subtraction analysis (Rorden & Karnath, 2004). To identify the white matter structures that are commonly damaged in patients who develop language representation predominantly in the right hemisphere but are typically spared in those with language representation predominantly in the left hemisphere, a first contrast was performed. We subtracted the superimposed lesions of those patients with predominant left hemisphere language representation (L > R) (Fig. 2A, 2nd row) from the overlap images of those with predominant language representation in the right hemisphere (R > L) (Fig. 2A, 1st row). Moreover, to identify the white matter structures that are commonly damaged in patients with disrupted crossed cortico-spinal projections to the paretic hand but are typically spared in patients with preserved hand motor projections, in a second contrast we subtracted the patients with preserved corticospinal hand motor projections Fig. 2A, 4th row) from the patients with disrupted cortico-spinal projections Fig. 2A, 3rd row). For both contrasts, an arbitrarily chosen threshold of 70% difference between groups was applied to define voxels that are critically involved, so that all voxels were marked with Number of patients in Group A with voxel lesioned Total number of patients in Group A Number of patients in Group B with voxel lesioned Total number of patients in Group B > 70% with Groups A and B being either patients with disrupted (A) versus preserved (B) hand motor projections, or with predominantly right-hemispheric (A) versus predominantly left-hemispheric (B) language activation. In our sample, this meant that, for the language contrast, all voxels were marked in which a lesion was present in 4/5 (80%) or 5/5 (100%) patients with right-hemispheric language organization (Group A), but in 0/3 (0%) of the patients with predominantly left-hemispheric language activation (Group B) resulting in supra-threshold differences of (80% 0% = 80%) or (100% 0% = 100%), respectively. For the hand motor contrast, all voxels were marked in which a lesion was present in 3/3 patients (100%) with disrupted hand motor projections (Group A 0 ), but in 0/4 (0%) or only 1/4 patients (25%) with preserved hand motor projections (Group B 0 ) resulting in supra-threshold differences of (100% 0% = 100%) or (100% 25% = 75%), respectively (Fig. 2B). To identify the anatomical relation of the affected white matter structures to the anatomical position of corticospinal/cortico-nuclear projections 1 originating in the primary motor cortex, we here used the new strategy of lesion analysis suggested by Papageorgiou and co-workers (in press). These authors combined established lesion subtraction techniques with the stereotaxic probabilistic cytoarchitectonic atlas, the latter developed by the Jülich group (Amunts & Zilles, 2001). In Fig. 3, we thus plotted the resulting subtraction images onto the stereotaxic probabilistic cytoarchitectonic map of the cortico-spinal/corticonuclear tract by Bürgel et al. (2006). This map illustrates the relative frequency with which the cortico-spinal/cortico-nuclear tract of ten normal human brains was present on a MNI reference brain in a voxel (e.g., a 50% value of the cortico-spinal tract in a certain voxel of the reference brain indicates that the fiber tract was present in that voxel in five out of ten brains). The probabilistic cytoarchitectonic map thus serves as a measure of intersubject variability for each voxel of the reference space. 3. Results We found the two centers of overlap in the white matter close to the left ventricle and clearly separated (Fig. 2). Plotting these centers onto the Jülich stereotaxic probabilistic cytoarchitectonic map revealed that both centers clearly affected the cortico-spinal/cortico-nuclear projection (Fig. 3). The area typically damaged in patients with disrupted cortico-spinal projections to the paretic hand (red area in Fig. 2B; pink contour in Fig. 3) extended from 1 In the paper by Bürgel and coworkers (Bürgel et al, 2006), all descending projections originating from primary motor cortex (M1) were identified as cortico-spinal projections. This nomenclature is somewhat misleading, since the approach these authors used also labelled motor projections from the M1 face representation (K. Amunts, personal communication). Since these fibers do not project to the spinal cord, but to the facial nerve nuclei, we used the term cortico-spinal/cortico-nuclear tract to refer to this structure.

4 180 M. Staudt et al. / Brain & Language 106 (2008) Fig. 1. Coronal reconstructions of the normalized 3D data sets depicting the extent of the lesion in each individual patient. MNI y-coordinates of the coronal sections are given. MNI coordinates (x, 21; y, 23; z, 31) over (x, 21; y, 20; z, 32) to (x, 20; y, 17; z, 35). In contrast, the white matter structure that was commonly damaged in patients who developed language representation predominantly in

5 M. Staudt et al. / Brain & Language 106 (2008) Fig. 2. (A) Conventional lesion overlay plots based on normalized T1-weighted MR images for the four groups of patients represented from top to bottom as follows: language representation predominantly in the right hemisphere, language representation predominantly in the left hemisphere, disrupted crossed cortico-spinal projections to the paretic hand, preserved cortico-spinal projections. The number of overlapping lesions is illustrated by different colours, coding increasing frequencies from violet (n = 1) to red (n = max). MNI y-coordinates of the coronal sections are given. L > R, language representation predominantly in the left hemisphere; R > L, language representation predominantly in the right hemisphere. Note that patient #4 could only be included in the analysis regarding language, but not regarding motor projections, since no TMS data could be obtained. (B) Overlay plots of the subtracted superimposed lesions of (i) patients showing language representation predominantly in the right hemisphere minus patients with language representation predominantly in the left hemisphere as well as (ii) patients with disrupted crossed cortico-spinal projections to the paretic hand minus patients with preserved cortico-spinal projections (for mathematical formula see Section 2). The differences between the respective groups in each voxel are indicated by colors: regions damaged at least 70% more frequently in patients who developed language representation in the right hemisphere are marked in dark green (80%) and light green (100%); regions damaged at least 70% more frequently in patients with a disruption of the cortico-spinal projections to the paretic hand are marked in dark red (75%) and light red (100%). the contralesional right hemisphere (green area in Fig. 2B; white contour in Fig. 3) was located more inferiorly (zdirection) and more anteriorly (y-direction). The area extended from MNI coordinates (x, 18; y, 22; z, 21) over (x, 17; y, 18; z, 21) and (x, 18; y, 13; z, 22) to (x, 8; y, 4; z, 16). Neither of the reverse contrasts, i.e., the contrast searching for brain areas typically damaged in patients with

6 182 M. Staudt et al. / Brain & Language 106 (2008) Fig. 3. Probabilistic cytoarchitectonic map of the cortico-spinal/cortico-nuclear tract by Bürgel and co-workers (2006). The color bar indicates the absolute frequency of voxels containing the cortico-spinal/cortico-nuclear tract from 1 (dark blue) individual brain to 10 (red, overlap of all ten [=Maximum] brains). The superimposed pink contour represents the center of the subtracted lesion overlap obtained in the present study for patients with disrupted crossed projections to the paretic hand (see red area in Fig. 2B). The superimposed white contour represents the center of the subtracted lesion overlap obtained in the present study for patients who developed language representation predominantly in the contralesional right hemisphere (see green area in Fig. 2B). preserved cortico-spinal hand motor tracts but not in patients with disrupted tracts, nor the contrast searching for brain areas typically damaged in patients with predominantly left-hemispheric language but not in patients with predominantly right-hemispheric language yielded any voxels fulfilling this condition. 4. Discussion Our present analysis revealed that topographically separate locations of early left periventricular brain lesions are critical for the disruption of cortico-spinal hand motor projections and for the induction of language reorganization in the contralesional right hemisphere. The congruency of the critical region for the integrity or disruption of hand motor projections with cytoarchitectonical probabilistic maps of the cortico-spinal tract suggest that the cortico-spinal system had already been established at the time of the insult (the early 3rd trimester of pregnancy; Krägeloh-Mann, 2004). This is consistent with observations by Eyre and colleagues (2000), who reported that, in the human fetus, outgrowing cortico-spinal projections have already reached their spinal target zones at the beginning of the 3rd trimester of pregnancy. In contrast, in the somatosensory system, we could recently demonstrate that outgrowing thalamo-cortical somatosensory projections can bypass a periventricular lesion along alternative routes, so that the new position of such pathways is no longer compatible with predictions from adult anatomy (Staudt et al, 2006). The white matter locus critically involved in inducing reorganization of language production in the contralesional right hemisphere is at least partially also located in the course of motor projections originating in the primary motor cortex, as determined from the present analysis combining lesion subtraction techniques with the Jülich stereotaxic probabilistic cytoarchitectonic atlas. Furthermore, this region is located anterior and inferior to the region critically involved in the integrity or disruption of hand motor projections, which is compatible with the position of facial motor projections relative to hand motor projections in the periventricular white matter: According to the homuncular organization of the primary motor cortex (Fig. 2b, left), the face is represented inferior and, due to the oblique course of the precentral gyrus, also anterior to the hand. In addition, in the somatotopic organization of the internal capsule, the facial motor projections pass through the genu and, thus, again anterior to the hand motor projections passing through the posterior limb. Together, these findings corroborate our previously published hypothesis (Staudt et al., 2001) that structural damage to facial motor projections plays an important role in inducing right-hemispheric language organization in such patients. This hypothesis is also in accordance with recent reports on a relevance of speech-related cortical motor areas already during the early phases of normal language development during the first year of life (Imada et al, 2006). One could further confirm this finding by testing the integrity or disruption of facial motor tracts in such patients using TMS. We refrained, however, from performing such measurements in our patients, since TMS of the face motor region often implies painful stimulation of temporal muscles. Future research on periventricular lesions might also include diffusion tensor imaging, to further enhance our understanding of the functional consequences of such lesions on specific white matter tracts. The observations of this study should not be mistaken as claiming that damage to facial motor projections is the

7 M. Staudt et al. / Brain & Language 106 (2008) only possible mechanism responsible for right-hemispheric organization of language in patients with left-sided periventricular lesions. Especially in the patients with larger lesions, it is quite evident that other white matter structures involved in language processing (e.g., the arcuate fasciculus) as well as subcortical grey matter structures such as the basal ganglia are at least partially damaged; furthermore, the periventricular lesions might also have had a negative influence on the microanatomical integrity of cortical language areas, even if this was not evident on structural MRI. Finally, all our patients were left-handed (most likely due to their right-sided hemiparesis), so that a potential influence of this pathological left-handedness on the organization of language cannot be excluded either. In conclusion, this study demonstrates that different functional consequences of early periventricular brain lesions can be attributed to topographically distinct loci of the white matter damage. Acknowledgments This work was supported by the Deutsche Forschungsgemeinschaft (SFB A4, C4, C5) and the European Union (PERACT- Marie Curie Early Stage Training MEST-CT ). References Amunts, K., & Zilles, K. (2001). Advances in cytoarchitectonic mapping of the human cerebral cortex. Neuroimaging Clinics of North America, 11, Ashwal, S., Russman, B. S., Blasco, B. A., Miller, G., Sandler, A., Shevell, M., et al. (2004). Practice parameter: Diagnostic assessment of the child with cerebral palsy. Neurology, 62, Banker, B. Q., & Larroche, J. C. (1962). Periventricular leukomalacia of infancy. Archives of Neurology, 7, Brett, M., Leff, A. P., Rorden, C., & Ashburner, J. (2001). Spatial normalization of brain images with focal lesions using cost function masking. NeuroImage, 14, Bürgel, U., Amunts, K., Hoemke, L., Mohlberg, H., Gilsbach, J. M., & Zilles, K. (2006). White matter fiber tracts of the human brain: Threedimensional mapping at microscopic resolution, topography and intersubject variability. NeuroImage, 29, Carr, L. J., Harrison, L. M., Evans, A. L., & Stephens, J. A. (1993). Patterns of central motor reorganization in hemiplegic cerebral palsy. Brain, 116, Eyre, J. A., Miller, S., Clowry, G. J., Conway, E. A., & Watts, C. (2000). Functional corticospinal projections are established prenatally in the human foetus permitting involvement in the development of spinal motor centres. Brain, 123, Imada, T., Zhang, Y., Cheour, M., Taulu, S., Ahonen, A., & Kuhl, P. K. (2006). Infant speech perception activates Broca s area: A developmental magnetoencephalography study. Neuroreport, 17, Krägeloh-Mann, I. (2004). Imaging of early brain injury and cortical plasticity. Experimental Neurology, 190, S84 S90. Lidzba, K., Staudt, M., Wilke, M., & Krägeloh-Mann, I. (2006). Visuospatial deficits in patients with early left-hemispheric lesions and functional reorganization of language: Consequence of lesion or reorganization? Neuropsychologia, 44, Maegaki, Y., Maeoka, Y., Ishii, S., Shiota, M., Takeuchi, A., Yoshino, K., et al. (1997). Mechanisms of central motor reorganization in pediatric hemiplegic patients. Neuropediatrics, 28, Muter, V., Taylor, S., & Vargha-Khadem, F. (1997). A longitudinal study of early intellectual development in hemiplegic children. Neuropsychologia, 35, Nezu, A., Kimura, S., Takeshita, S., & Tanaka, M. (1999). Functional recovery in hemiplegic cerebral palsy: Ipsilateral electromyographic responses to focal transcranial magnetic stimulation. Brain and Development, 21, Papageorgiou, E., Ticini, L. F., Hardiess, G., Schaeffel, F., Wiethoelter, H., Mallot, H., Bahlo, S., Wilhelm, B., Vonthein, R., Schiefer, U., & Karnath, H. O. (in press). Anatomy of the pupillary light reflex pathway: cytoarchitectonic probabilistic maps in hemianopic patients. Neurology. Rorden, C., & Brett, M. (2000). Stereotaxic display of brain lesions. Behavioral Neurology, 12, Rorden, C., & Karnath, H.-O. (2004). Using human brain lesions to infer function: a relic from a past era in the fmri age? Nature Reviews Neuroscience, 5, Staudt, M., Braun, C., Gerloff, C., Erb, M., Grodd, W., & Krägeloh- Mann, I. (2006). Developing somatosensory projections bypass periventricular brain lesions. Neurology, 67, Staudt, M., Gerloff, C., Grodd, W., Holthausen, H., Niemann, G., & Krägeloh-Mann, I. (2004). Reorganization in congenital hemiparesis acquired at different gestational ages. Annals of Neurology, 56, Staudt, M., Grodd, W., Gerloff, C., Erb, M., Stitz, J., & Krägeloh-Mann, I. (2002a). Two types of ipsilateral reorganization in congenital hemiparesis: A TMS and fmri study. Brain, 125, Staudt, M., Grodd, W., Niemann, G., Wildgruber, D., Erb, M., & Krägeloh-Mann, I. (2001). Early left periventricular brain lesions induce right hemispheric organization of speech. Neurology, 57, Staudt, M., Lidzba, K., Grodd, W., Wildgruber, D., Erb, M., & Krägeloh- Mann, I. (2002b). Right hemispheric organization of language following early left-sided brain lesions: Functional MRI topography. NeuroImage, 16, Tewes, U. (1994). Hamburg-Wechsler Intelligenztest für Erwachsene. Revision 1991 (2nd ed.). Huber, Bern. Wildgruber, D., Ackermann, H., Klose, U., Kardatzki, B., & Grodd, W. (1996). Functional lateralization of speech production at primary motor cortex: A fmri study. Neuroreport, 7,