This article was downloaded by: [University of West Florida] On: 09 October 2014, At: 22:02 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Neurocase: The Neural Basis of Cognition Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/nncs20 Recovery of Semantic Word Processing in Transcortical Sensory Aphasia: a Functional Magnetic Resonance Imaging Study Roland Zahn , Walter Huber , Eva Drews , Karsten Specht , Stefan Kemeny , Wolfgang Reith , Klaus Willmes & Michael Schwarz Published online: 09 Aug 2010. To cite this article: Roland Zahn , Walter Huber , Eva Drews , Karsten Specht , Stefan Kemeny , Wolfgang Reith , Klaus Willmes & Michael Schwarz (2002) Recovery of Semantic Word Processing in Transcortical Sensory Aphasia: a Functional Magnetic Resonance Imaging Study, Neurocase: The Neural Basis of Cognition, 8:5, 376-386 To link to this article: http://dx.doi.org/10.1076/neur.8.4.376.16189 PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions Neurocase (2002) Vol. 8, pp. 376–386 © Oxford University Press 2002 Recovery of Semantic Word Processing in Transcortical Sensory Aphasia: a Functional Magnetic Resonance Imaging Study Roland Zahn1,2,5, Walter Huber2, Eva Drews1,2, Karsten Specht2,3, Stefan Kemeny3, Wolfgang Reith3, Klaus Willmes4 and Michael Schwarz1 1Department of Neurology, 2Neurolinguistics Section, 3Department of Neuroradiology, 4Neuropsychology Section, University of Downloaded by [University of West Florida] at 22:02 09 October 2014 Technology, RWTH Aachen, Pauwelsstrasse 30, D-52057 Aachen and 5Department of Psychiatry, University of Freiburg, Hauptstrasse 5, D-79104 Freiburg, Germany Abstract In a previous functional magnetic resonance imaging (fMRI) study with normal subjects, we demonstrated regions related to conceptual-semantic word processing around the first frontal sulcus (BA 9) and the posterior parietal lobe (BA 7/40) in agreement with several previous reports. We had the possibility, using the same fMRI paradigm, to study two consecutive cases with left middle cerebral artery (MCA) infarction (RC and HP) and lesions affecting either solely the pre-frontal (HP) or both the pre-frontal and posterior parietal part of the network activated in normal subjects (RC). Both patients showed transcortical sensory aphasia (TSA) on acute assessment. This contradicts classical disconnection accounts of the syndrome stating intact conceptual representations in TSA. Their recovery of language comprehension was associated with activation of a left hemispheric network. Mainly activations of left perilesional prefrontal regions (RC), left Wernicke’s area (RC and HP) or the left posterior middle and inferior temporal cortex (HP) were demonstrated in the TSA patients. The latter findings suggest that in our cases of TSA functional take-over has occurred in regions with related functions (‘redundancy recovery’) rather than in previously unrelated areas (‘vicarious functioning’). Our data support distributed models of conceptual-semantic word processing and multiple left hemispheric representations of closely related functions. Introduction The dissociation of impaired auditory word comprehension and relatively intact repetition termed transcortical sensory aphasia (TSA) has been taken as evidence for an anatomical disconnection of the auditory word–images centre, i.e. Wernicke’s area, from a diffuse posterior ‘ideational field’ (Begriffsfeld) (Lichtheim, 1885). Disconnection mechanisms have been emphasized in neoclassical models of language functions (Geschwind, 1965). Some modern authors view lesions in TSA as representing partial damage to a distributed semantic system itself, rather than a disconnection of intact centres for conceptual and lexical processing (Alexander et al., 1989). Although functional imaging studies have corroborated the view of a distributed semantic system in posterior extrasylvian areas (Démonet et al., 1992; Warburton et al., 1996; Binder et al., 1997; Vandenberghe et al., 1996), they seem to contradict lesion studies of standard aphasic syndromes, including Wernicke’s aphasia, where the most consistent overlap of lesions associated with persistent impairments of comprehension was to be found in the posterior perisylvian (Wernicke’s area), but not the extrasylvian cortex (Willmes and Poeck, 1993). An explanation for this contradiction has been offered by postulating a disconnection of damaged lexical representations in Wernicke’s area from spared left extrasylvian conceptual representations (Warburton et al., 1996) or from bilaterally represented conceptual categories (Gainotti et al., 1995) in aphasia. These modern disconnection accounts are compatible with recovery of language functions within a localizationist model. The remarkable ability to recover from functional impairments, however, has been challenging localizationist approaches in general [reviewed by Phillips et al. (1984)]. Several alternative theories have been put forward to account for recovery of function after brain injury [reviewed in Marshall (1984) and Finger and Almli (1985)]. Theories of functional compensation (Finger and Almli, 1985) or so-called redundancy recovery (Marshall, 1984) contradict a strict and uniform localization of cognitive functions. These theories presuppose distributed representa- Correspondence to: Roland Zahn, Department of Psychiatry and Psychotherapy, University of Freiburg, Hauptstrasse 5, D-79104 Freiburg, Germany. Tel: ⫹49 761 270 6642; Fax: ⫹49 761 270 6619; e-mail: roland_zahn@psyallg.ukl.uni-freiburg.de Downloaded by [University of West Florida] at 22:02 09 October 2014 Recovery of semantic word processing in TSA 377 tions of closely related or identical functions. For language the existence of such multiple representations of functions has been concluded from the equivocal relationship between standard aphasic syndromes and the localization of corresponding lesions (Willmes and Poeck, 1993). Whereas structure–function relationships have been primarily established by looking for the association of deficit symptoms or recovered functions with certain lesions, functional activation methods offer a means to study more directly the functional anatomy of these recovered or retained functions in neuropsychological patients [for a review of theories of deficit and compensation in aphasia see Springer et al. (2000)]. We were interested to test the predictions of theories of recovery applied to recovered semantic word processing in two consecutive cases of TSA after left middle cerebral artery (MCA) infarction, using the Aachen Aphasia Test (AAT). To our knowledge, this is the first functional activation study measuring semantic processing in TSA after a left hemispheric stroke. In a previous functional magnetic resonance imaging (fMRI) study with normal subjects, including elderly persons, we were able to activate left pre-frontal (BA 9) and posterior parietal (BA 7/40) areas in accordance with previous studies [reviewed in Price (1998)] as part of the conceptual-semantic network using a semantic monitoring task on auditory words versus auditory lexical decision with non-word distractors (Zahn et al., 2000). (1) According to the disconnection theory, areas normally involved in lexical or conceptual processing should be spared in TSA and recovery of auditory comprehension should be associated with a re-establishment of the connections between these functionally reactivated areas. (2) Vicarious functioning (Marshall, 1984) can only be assumed when areas are activated after recovery which never show detectable activity in normal subjects during the measured task. (3) Recovery by functional compensation (Finger and Almli, 1985) or redundancy recovery (Marshall, 1984) within a distributed network of semantic functions is expected to occur through activation of spared parts of a partially damaged semantic system. (4) The left-to-right transfer hypothesis predicts that recovery of auditory comprehension is mediated by activation of right hemisphere homologues of language areas that might have had latent language capabilities prior to disease [reviewed in Kinsbourne (1998)]. Case histories TSA was diagnosed according to the operational criteria given by the AAT (Huber et al., 1984): fluent speech with no difficulties in articulation, repetition subtest ⬎ percentile rank 50 and repetition subtest performance 20 percentile ranks above comprehension subtest, naming subtest and Token Test performance. HP strictly and RC essentially fulfilled these criteria (Table 1). RC RC was a 65-year-old right-handed former mechanical engineer who suffered from a cardio-embolic left hemispheric stroke in January 1998. One day post-onset, a computed tomography (CT) scan revealed two hypodensities in the frontal and parietal MCA territory. On neurological examination a mild right facial paresis and a latent paresis of the right arm were noted. Fluent spontaneous speech with severe language comprehension deficits were apparent. A comparison of two conventional MRI examinations using coronal and axial Fluid Attenuated Inversion Recovery (FLAIR) sequences performed after 10 days and 5 months post-stroke revealed only a slight retraction of the initial hyperintense frontal and parietal lesions. T1-weighted three-dimensional MRI scanning after 5 months indicated an almost complete hypointense demarcation of these areas. Extra- and transcranial Doppler sonography were normal. An arrhythmia absoluta pointed to a cardio-embolic event as the most probable cause. Aphasia was assessed using the AAT (Huber et al., 1984) on three occasions during the acute and post-acute phases after stroke (Table 1). In the acute phase, spontaneous speech was fluent with no difficulty initiating speech and with normal articulation. Severe word-finding difficulties, semantic paraphasias and many circumlocutions were observed, which were often repeated stereotypically. Syntactic constructions were complex and paragrammatic. Initially there were also many phonemic paraphasias and searching, but these symptoms almost disappeared during the first month, whereas paragrammatic and semantic abnormalities recovered more slowly. The results of the first AAT examination confirmed TSA with repetition being significantly better than all of the other subtests. In the last AAT examination at 5 months, the time of fMRI scanning, the ceiling level had been reached. HP HP was a 69-year-old right-handed former high school teacher who suffered from a cardio-embolic left hemispheric stroke in October 1998. One day post-onset a CT scan revealed a left frontal hypodensity within the MCA territory. On neurological examination no focal deficits were noted. Fluent spontaneous speech with severe language comprehension deficits were apparent. A comparison of two conventional MRI examinations using coronal and axial FLAIR sequences performed after 12 days and 6 months post-stroke revealed only a slight retraction of the initial hyperintense frontal lesion. T1-weighted three-dimensional MRI scanning after 6 months indicated an almost complete hypointense demarcation of this area. Extra- and transcranial Doppler sonography was normal. An intermittent arrhythmia absoluta pointed to a cardio-embolic event as the most probable cause. In the acute phase, spontaneous speech was fluent with no difficulty initiating speech and with normal articulation. Severe word-finding difficulties, semantic paraphasias and many circumlocutions were observed, which were often repeated stereotypically. Syntactic constructions were complex and paragrammatic. From the beginning there were no phonemic paraphasias. 378 R. Zahn et al. Table 1. Results of the Aachen Aphasia Test examinations Downloaded by [University of West Florida] at 22:02 09 October 2014 Spontaneous speech rating scales Subtests (percentiles) Time post-stroke COM ART AUT SEM PHO SYN TT REP WRIT NAME COMP RC 12th day 10th week 5 months 3 4 5 5 5 5 3 4 4 3 4 4 3 4 5 3 4 4 31 91** 97 70 94** 96 53 99** 99 52 91** 93 52 88(**) 95 HP 10th day 12th week 6 months 2 4** 5 5 5 5 2 5** 5 2 4** 5 5 5 5 3 5** 5 33 83** 94 88 99 97 – 99 100 41 88** 94 21 59** 70 The Aachen Aphasia Test consists of six six-point spontaneous speech rating scales with 0 ⫽ most deviant and 5 ⫽ normal, and five subtests. Subtest scores are reported as percentile ranks based on norms for the aphasic population (n ⫽ 376). COM, communicative behaviour; ART, articulation and prosody; AUT, automized and stereotypical speech elements; SEM, semantic structure; PHO, phonological structure; SYN, syntactic structure; TT, Token Test; REP, repetition; WRIT, written language; NAME, confrontation naming; COMP, comprehension, in which different units (phonemes, monosyllabic and polysyllabic nouns, sentences) and linguistic rules are incorporated. **, Significant differences in comparison with the subtest performance in the preceding examination (α ⫽ 10 %). (**), Differences that reach 75% of the critical value for a significant difference. Severity of impairment in a subtest is operationally defined by Stanine norms (1–9). For repetition, percentile ranks (PR) and corresponding Stanine norms (ST): severe impairment: ST 1–3 ⫽ PR 1–25; moderate impairment: ST 4–5 ⫽ PR 26–64; mild impairment: ST 6–7 ⫽ PR 65–89; residual or no impairment: ST 8–9 ⫽ PR 90–100. In the first AAT examination, 10 days post-stroke, HP performed significantly better in the repetition subtest compared with all of the other subtests (Table 1). In the last AAT examination at 6 months, the time of fMRI scanning, the ceiling level had been reached for all subtests, but comprehension was still impaired. Materials and methods Subjects RC and HP and healthy male subjects (n ⫽ 14; 23–65 years), strongly right handed, gave informed consent. fMRI study The scanning procedure was identical to the normal group study, where we used an additional condition, described elsewhere (Zahn et al., 2000). In this study we focused on two conditions. In the lexical condition [L] an auditory lexical decision task was used, where reversed words served as phonotactically illegal non-word distractors. The use of non-words instead of pseudowords minimizes segmental phonological and automatic lexical-semantic processing. In addition, the target/distracter ratio was kept low at 1/3 across all conditions to reduce further automatic activation of word meaning for targets. The task was expected to engage matching processes involved in lexical access, in addition to pre-lexical processing. In the semantic condition [S] animals had to be discriminated from names of other natural kind, thus requiring a superordinate category decision. Within the chosen model (Hillert, 1999), we expected mainly conceptualsemantic in addition to lexical-semantic and word form processing to be activated. By ‘conceptual-semantic’ we denote an analysis of meaning that goes beyond the meaning assigned to a lexical entry, by requiring (de-) compositions like superordinate category judgements (Hillert, 1999). All stimuli were matched for stimulus duration and sound amplitude, word frequency and number of syllables of German. Digital stimuli were auditorily presented (1/3 s) at 85 dB. During the activation periods, the subjects had to monitor for targets according to the criterion given before each session and respond with a button press. Each experimental condition consisted of one session with three 88-s cycles alternating between rest (off) and activation (on). Whole brain fMRI was performed on a 1.5T Philips ACS NT Gyroscan with a standard head coil and echo-planar imaging [repetition time (RT) 4000 ms; echo time (ET) 40 ms; field angle (FA) 40°; matrix 64 ⫻ 64; field of view (FOV) 220 ⫻ 170 mm; 15 contiguous 7 mm slices parallel to the AC–PC line]. After the functional scans, a high-resolution volumetric anatomic scan was acquired for three-dimensional rendering of the brain (3D FFE sequence with RT 30 ms; ET 4.5 ms; FA 30°; matrix 256 ⫻ 256; FOV 220 ⫻ 176 mm; 90 contiguous 2 mm thick slices). The data were analysed using SPM96 software (Wellcome Department of Cognitive Neurology, London, UK). The functional images were spatially realigned, normalized into standard stereotaxic space, resliced with a voxel size of 4 ⫻ 4 ⫻ 7 mm3 and smoothed (8 ⫻ 8 ⫻ 14 mm3 kernel) (Friston et al., 1995). The first scans from each on and off period were discarded. In the present study we focused on the following complex contrast (R ⫽ rest): interaction between semantic and lexical task indicating areas associated with the processing of word meaning, mainly conceptual-semantic word processing: [S–R]–[L–R], abbreviated [S–L]. We employed conjunction over subjects for the normal group (contrasts of individual subjects separately modelled within one design matrix) showing activation maps of voxels Downloaded by [University of West Florida] at 22:02 09 October 2014 Recovery of semantic word processing in TSA 379 consistently activated in all subjects (Price and Friston, 1997). Maxima exceeding an individual voxel level of P ⫽ 0.001 (z ⫽ 3.09) and a cluster size ⬎ 16 voxels (P ⬍ 0.05 corrected) are displayed in Fig. 1 and Table 2. Because we were interested in the interindividual variability in the normal subjects, we also analysed normal single cases (each subject modelled again within a single design matrix) looking at contrast [S–L] in each subject in five left hemispheric regions of interest (ROIs) defined by Talairach coordinates (see Table 3): (1) posterior middle pre-frontal, (2) inferior prefrontal cortex, (3) posterior parietal cortex, (4) posterior superior temporal gyrus and (5) posterior middle/inferior temporal gyri. Maxima exceeding P ⫽ 0.01 (z ⬎ 2.33) are reported within these regions in Table 3. ROIs (1) and (3) were chosen based on the normal group study and ROIs (2), (4) and (5) for comparison with the main results of the single case analysis in RC and HP (see below). We performed a single case analysis for RC’s and HP’s data. Because of spatial distortions occurring during spatial normalization of brains with large lesions, which exceed those tolerable in brains without lesions, we performed a statistical analysis based on the non-normalized but smoothed functional images. Activation maps were then projected on to the individual three-dimensional T1-weighted anatomical image that had been co-registered with the functional images by using the anatomical T1-weighted reference image (Pietrzyk et al., 1994). Localization of activations was determined in axial, coronal and sagittal slices by an experienced neuroradiologist (WR) using anatomical landmarks to identify the activated structures in an anatomical reference atlas (Mai et al., 1998). Results Normal subjects’ and patients’ performances in the fMRI tasks There were no significant differences (Wilcoxon test, P ⫽ 0.10) between the tasks in the normal group with respect to d’ values (Table 4). Performances of RC and HP in the semantic task were subnormal but significantly above a chance level response (d’ around 2). The relative d’ difference in performance between tasks was within the normal range. Structural MRI RC. The left hemispheric lesion involved the posterior perisylvian inferior parietal lobule, extending into the parietooccipital junction as well as parts of the inferior and large parts of the left middle frontal gyrus. The pars opercularis of the inferior frontal gyrus (posterior frontal operculum), the superior frontal gyrus and the temporal lobe were spared. HP. The left hemispheric lesion involved parts of the inferior and large parts of the left middle frontal gyrus. The pars opercularis of the inferior frontal gyrus (posterior frontal operculum) and the superior frontal gyrus were spared. The lesion was strikingly similar to RC’s frontal lesion. fMRI study Semantic versus lexical condition [S–L]. Figure 1 shows T1-weighted axial magnetic resonance slices parallel to the AC–PC line through RC’s and HP⬘s brains with superimposed activation maps obtained from spatially non-normalized functional images. For comparison, axial slices through a standard brain (Evans et al., 1994) at comparable levels along the z-axis with normalized normal group data are displayed (conjunction over subjects z ⬎ 3.09, P ⬍ 0.05 corrected). In Table 2 the results of the normal group (conjunction over subjects) are reported and in Table 3 the results of the ROI single normal subject analysis using normalized data are reported. Table 5 gives statistical information for RC and HP and the anatomical localization of activations. The numbers in the text and in Table 5 refer to the numbered activation peaks in Fig. 1. Normal subjects Activations in the contrast semantic versus lexical task were viewed as conceptual-semantic processing areas. For the group (conjunction over subjects) these included parts of the posterior left middle frontal gyrus (BA 9, near BA 45) and posterior parietal lobe (BA 7/40) as seen in Fig. 1 and Table 2. The ROI-based analysis of single normal subjects (Table 3) showed that 12 of 14 single subjects displayed a posterior inferior or middle pre-frontal activation (BA 9, 46/9, 45, 47) at P ⫽ 0.01. Posterior parietal activations reached significance in eight of 14 single subjects, scattered over BA 7/19, 7/40 and BA 39. Posterior temporal activations, although not significant in the group study, were found in eight subjects; these were located in the posterior inferior or middle temporal gyrus (BA 20/21/37). Only one subject displayed an activation within the posterior superior temporal gyrus (BA 22). Single subject analysis for RC and HP based on nonnormalized data. Inferior and middle frontal activations occurred perilesionally (3, 5, 6, 9). RC showed a perilesional activation in the parietal extension of the superior temporal gyrus (4) as well as left lateralized thalamic (2), left precuneus activations (7) and a left pre-central signal increase (8). There was a single highly significant cluster of activation within the left posterior superior, middle and inferior temporal gyrus. Thus, both patients activated parts of Wernicke’s area, i.e. the posterior superior temporal gyrus. Other activations were not observed, even on a more liberal level of significance (z ⬎ 2.33, P ⫽ 0.01 uncorrected). Discussion Functional recovery of semantic single word processing in RC and HP In the present investigation we were interested in the functional anatomy of recovered conceptual-semantic word Downloaded by [University of West Florida] at 22:02 09 October 2014 380 R. Zahn et al. Fig. 1. Spatially non-normalized functional data for RC and HP showing areas of significant activation specific to the semantic relative to the lexical task [S–R]–[L–R] are displayed in the second, third and fifth, sixth rows. The activations are superimposed on eight slices (7 mm interslice interval) parallel to the AC–PC (anterior–posterior commissure) line from the patients’ non-normalized anatomical T1-weighted three-dimensional magnetic resonance image. Probability maps of the conjunction analysis for the normal group at comparable levels along the z-axis with normalized normal group data are shown in the first and fourth rows. The normal group activations (conjunction over subjects) are superimposed on eight slices parallel to the AC–PC line (z-coordinates ⫽ –14, –7, 0, 14, 21, 28, 35, 42 mm) from a standard brain (Evans et al., 1994). Activations reaching an individual voxel significance level threshold of P ⫽ 0.001 and a corrected cluster level ⬍ 0.05 for the normal group and a voxel level of P ⬍ 0.01 uncorrected, 10 voxels, for the patients are displayed. Recovery of semantic word processing in TSA 381 Table 2. Increases in brain activity, semantic versus lexical task, normal group conjunction analysis (n ⫽ 14) Area Maxima (x, y, z) BA z score Cluster P (corrected) Talairach ROI for comparison with single subject data Anatomical description of ROI (x–x), (y–y), (z–z) Middle frontal gyrus –40, 24, 28 9 4.18 0.032 x: (–34 to –60), y: (0 to ⫹45), z: (⫹9 to ⫹40) Inferior and middle frontal gyrus (BA 44/45/9) 7 Posterior parietal cortex/ parieto-occipital junction –32, –60, 49 7/40 4.59 0.028 x: (–30 to –55), y: (–55 to –70), z: (⫹20 to ⫹50) Posterior parietal cortex (BA 7/40/39) 8 Na aN ⫽ number of single subjects with activation within ROI at P ⫽ 0.01 (uncorrected). Downloaded by [University of West Florida] at 22:02 09 October 2014 All areas exceeding voxel level: z ⬎ 3.09 and P ⬍ 0.001, corrected cluster level P ⬍ 0.05 (⬎ 16 voxels). ROI, region of interest. processing in two cases of TSA after left MCA infarction (RC and HP) using fMRI. By discussing several different mechanisms of recovery we hope to contribute to the debate about localized versus distributed models and the role of the pre-frontal cortex in semantic word processing. Aphasiological profiles and lesions In accordance with the disconnection theory of TSA (Lichtheim, 1885), Wernicke’s and Broca’s areas were largely spared in RC and HP. Their lesions, however, involved parts of the putative conceptual-semantic word processing system proper, as posterior parietal and pre-frontal areas, consistently activated in our normal group as well as in several previous studies (Démonet et al., 1992; Warburton et al., 1996; Binder et al., 1997), were damaged. Consequently, intact conceptual knowledge after a lesion within these areas is incompatible with a simple anatomical disconnection of pathways between intact conceptual knowledge and lexical representations. One could argue that our semantic monitoring task was inadequate to represent conceptual-semantic word processing in general. But even if one admits the limitations of our activation paradigm, that mainly requires superordinate judgement on animate concepts and cannot encompass conceptual processing in general, the recovery of the kind of conceptualsemantic word processing involved in the task was also relevant to the clinically observed recovery of initially impaired auditory single word comprehension (as measured by the AAT). Even if the task can only activate part of the conceptual network, a demonstration of lesions within parts of the network that have led to impairments argues against disconnection theory. TSA has typically been observed after lesions damaging the temporo-occipital or parieto-occipital junction (Kertesz et al., 1982; Alexander et al., 1989). TSA after frontal lesions were repeatedly discarded as atypical [for a recent review see Berthier (2001)]. Berthier (2001) discussed that primarily atypical language organization (e.g. more bilateral language representation) might contribute to the occurrence of TSA after a left frontal lesion. The results of our functional activation study in both cases make a primarily bilateral organization of language improbable. Our conclusions might not apply to all cases of TSA with other types of lesions. The finding of an overlap of lesions within the pre-frontal cortex in two cases with TSA taken together with the previously reported cases, however, can be viewed as clinical evidence for a contribution of the pre-frontal cortex to word comprehension. Pre-frontal activations, that were consistently found in a variety of different semantic tasks with functional imaging [discussed and reviewed in Gabrieli et al. (1998)], seem to support this conclusion. Because pre-frontal activations in receptive semantic tasks seemed to contradict classical theories of language localization, one was cautious to argue that pre-frontal regions are really involved in word comprehension proper and argued for semantic search, selection or executive demands (Thompson-Schill et al., 1997; Price et al., 1999). But in our view, further clinical evidence for a frontal contribution to word comprehension comes from the fact that the observed preservation of word comprehension in Broca’s aphasia (most consistently involving frontal cortex; Willmes and Poeck, 1993) is a result of recovery from initially severe impairments within months or years (Mohr et al., 1978). Price et al. (1999) concluded from their positron emission tomography (PET) study in a patient (SW) with a left inferior pre-frontal lesion without impairment, in a semantic similarity task 4 years post-stroke and showing no perilesional activations, that the inferior frontal activation observed in normal subjects was not a necessary component of the semantic system. Our data are compatible with a modification of this view in showing that independence of the semantic system from the pre-frontal cortex can occur as a result of recovery. One reason for the rarity of reports of pre-frontal lesions leading to TSA might be that most lesion studies have been conducted at a chronic stage. An objection could be made to our interpretation of comprehension deficits in frontal lesion patients, that it is not the frontal lesion itself in Broca’s aphasia or in our TSA cases producing the impairment, but that effects on areas distant to the lesion (diaschisis), for example posterior perisylvian areas, lead to the deficits. Cappa et al. (1997) suggested that diaschisis might be the predominant mechanism of recovery in acute aphasia. The classical theory of diaschisis within a localizationist framework states that 382 R. Zahn et al. Table 3. Increases in brain activity, semantic versus lexical task. Region of interest (ROI) analysis of single normal subject Posterior pre-frontal x: (–34 to –60) y: (0 to ⫹45) z: (⫹9 to ⫹40) Inferior pre-frontal x: (–34 to –60) y: (0 to ⫹45) z: (–14 to ⫹8) Posterior parietal x: (–30 to –55) y: (–55 to –70) z: (⫹20 to ⫹50) Posterior STG x: (–50 to –60) y: (–35 to –55) z: (⫹10 to ⫹20) Posterior MTG, posterior ITG x: (–32 to –60) y: (–35 to –68) z: (–21 to ⫹7) 1 2 No activation No activation No activation No activation No activation 3 MFG BA 9 –40, 20, 35 z ⫽ 4.53 No activation IFG BA 47 –28, 36, –7 z ⫽ 3.46 No activation G.ang. / LPI BA 39 –44, –68, 28 z ⫽ 4.55 No activation No activation ITG BA 37 –60, –68, 7 z ⫽ 2.37 No activation 4 5 No activation No activation No activation MFG BA 46/9 –32, 28, 28 z ⫽ 3.81 No activation LPI BA 7 –24, –56, 56 z ⫽ 3.76 GTM/G.ang. BA 39 –32, –72, 14 z ⫽ 2.97 No activation No activation 6 No activation IFG BA 47 –40, 40, –14 z ⫽ 3.82 No activation No activation 7 MFG BA 9/46 –60, 32, 28 z ⫽ 3.30 IFG BA 44 –36, 8, 35 z ⫽ 2.99 No activation No activation No activation ITG BA 37 –48, –52, –21 z ⫽ 4.13 MTG BA 21 –44, –56, 7 z ⫽ 3.12 MTG BA 37 –60, –56, –7 z ⫽ 4.30 IFG BA 47 –44, 20, –14 z ⫽ 3.34 IFG BA 47 –44, 36, –7 z ⫽ 3.59 No activation No activation GTM / G. ang. BA 39 –40, –72, 21 z ⫽ 3.04 No activation No activation STG BA 22 –56, –52, 21 z ⫽ 2.35 No activation Downloaded by [University of West Florida] at 22:02 09 October 2014 Case 8 9 IFG/MFG BA 45/9 –36, 20, 21 z ⫽ 4.05 MFG BA 46/9 –40, 28, 28 z ⫽ 3.51 10 11 12 13 14 Na MFG 9/46 –28, 36, 28 z ⫽ 3.30 MFG/IFG BA 9/46 –36, 40, 21 z ⫽ 3.17 IFG/MFG BA 9 –52, 24, 40 z ⫽ 3.16 7 IFG 45 –40, 40, 0 z ⫽ 4.18 MFG 46/9 –44, 44, 0 z ⫽ 4.15 No activation No activation No activation 6 LPI BA 7/19 –36, –72, 42 z ⫽ 3.94 LPI BA 40 –64, –44, 35 z ⫽ 3.34 LPI BA 7/40 –40, –44, 42 z ⫽ 2.44 LPI BA 40 –48, –48, 49 z ⫽ 3.12 8 ITG BA 20 –68, –40, –7 z ⫽ 3.27 MTG BA 21 –36, –52, 7 z ⫽ 4.07 No activation ITG BA 37 –32, –60, –14 z ⫽ 4.38 ITG BA 20 –52, –40, –21 z ⫽ 4.47 No activation No activation No activation No activation MTG/ITG BA 21/37 –56, –48, 0 z ⫽ 3.84 8 1 aNumber of subjects with activation at P ⫽ 0.01. All areas exceeding a voxel level significance threshold of z ⫽ 2.33, P ⫽ 0.01 (uncorrected) within the defined ROI are reported. ITG, inferior temporal gyrus; MTG, middle temporal gyrus; STG, superior temporal gyrus; IFG, inferior frontal gyrus; MFG, middle frontal gyrus; LPI, inferior parietal lobule; G. ang., angular gyrus. Coordinates in italic ⫽ maximum out of ROI range, but activated cluster extends into ROI. reversible deficits are due to functional deactivation of structurally unaffected areas distant to the lesion (Feeney and Baron, 1986). The areas lesioned in RC and HP, however, were likely to have contributed to their initially impaired semantic word processing prior to stroke, based on the evidence of our, and previous, functional activation studies. This would contradict the classical concept of diaschisis. A less strict interpretation of diaschisis, however, would be compatible with our data. When multiple representations of closely related functions are assumed, regression of diaschisis Recovery of semantic word processing in TSA 383 Table 4. d’ values for the tasks in each subject Downloaded by [University of West Florida] at 22:02 09 October 2014 Normal subjects 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Mean Range RC HP Age 23 25 25 26 25 24 27 65 60 58 54 52 53 50 Sex m m m m m m m m m m m m m m m m d’ difference (d’ semantic– d’ lexical) –3.6 2.0 0 0 –1.9 1.6 1.3 –1.8 –1.5 0 –1.6 0 –3.9 3.6 –0.4 (–3.9–3.6) –3.2 –0.7 Table 5. Increases in brain activity, semantic versus lexical task, in RC and HP, non-normalized data Lexical task d’ 6.8 4.9 6.8 6.8 4.8 5.2 3.2 4.3 4.1 2.9 6.8 6.8 6.8 3.3 5.2 (2.9–6.8) 5.2 2.6a Semantic task d’ 3.2 6.8 6.8 6.8 3.0 6.8 4.6 2.4 2.6 2.9 5.2 6.8 2.9 6.8 4.8 (2.4–6.8) 2.0a 1.9a Wilcoxon sum rank test: no significant differences between the conditions for the normal subjects (P ⬍ 0.10) for d’. aOut of normal range. could explain the time a partially redundant system needs to compensate for damaged parts by areas left structurally unaffected. Functional activations associated with recovery of conceptual-semantic word processing in RC and HP The two most important results of our functional activation study in RC and HP were: (1) that at the time of recovery of auditory comprehension (around 6 months post-stroke), left posterior perisylvian areas were functionally active, furthermore, specifically active during conceptual-semantic versus lexical decision [S–L] and (2) that there was no relevant activation of right hemispheric cortex associated with semantic word processing. Activation of the posterior superior temporal gyrus (Wernicke’s area) in both patients was a surprising finding, as Wernicke’s area was not activated in our normal group and even on a single case basis there was an activation only in one subject (Table 3). Although Wernicke’s area was not activated in our study, a variety of evidence from other functional activation (Warburton et al., 1996) as well as clinical studies (Naeser et al., 1987; Willmes and Poeck, 1993) leaves little doubt that the posterior superior temporal gyrus is involved in auditory comprehension. The exact role of Wernicke’s area, however, is still under debate. In a model proposed by Mesulam (1998) it is suggested to function as a ‘transmodal gateway’ that mediates between word form and distributed semantic representations. A TSA-like picture has recently been reported after electrocortical stimulation within the left posterior middle temporal gyrus and superior temporal gyrus. The authors suggest disconnection of phono- Number corresponding to Fig. 1 RC 1 2 3 4 5 6 7 8 9 HP 1 Area z score Cluster size at P ⫽ 0.01 Right periventricular Left thalamus Left anterior inferior frontal gyrus Left posterior superior temporal gyrus, parietal extension Left middle frontal gyrus Left anterior middle frontal gyrus Left pre-cuneus Left pre-central Left posterior middle frontal gyrus 4.52 3.84 4.40 19 26 12 4.32 23 4.21 4.86 10 98 4.24 4.43 4.17 13 24 20 Left posterior superior, 4.50 middle and inferior temporal gyrus 54 All areas exceeding a voxel level significance threshold of z ⫽ 3.09, P ⫽ 0.001 (uncorrected) and a cluster size 艌 10 are reported for RC and HP. logy from lexical-semantics by lesions in Wernicke’s area (Boatman et al., 2000) compatible with Mesulam’s model. As the posterior superior temporal gyrus probably has related functions to the semantic processing measured in our task and is at least activated in one normal subject, one cannot speak of a vicarious activation. Activation of the spared Wernicke’s area in RC and HP demonstrates that it is functionally intact. This points to a compensatory role in the recovery of semantic word processing, either as part of a distributed conceptual system or as a correlate of lexical processes that could partially compensate for damaged conceptual-semantic functions. The demonstration of a preserved left hemispheric dominance for conceptual-semantic word processing in our patients conflicts with the left-to-right transfer hypothesis of language recovery supported by the studies of Weiller et al. (1995), Buckner et al. (1996), Ohyama et al. (1996), Cao et al. (1999), Musso et al. (1999) and Thulborn et al. (1999) or a primarily bilateral representation of word processing in some individuals (Frackowiak, 1997) as correlates of good recovery. Other functional activation studies have stressed the importance of left hemispheric compensation in the recovery from aphasia (Karbe et al., 1998; Heiss et al., 1999; Miura et al., 1999; Warburton et al., 1999). It is not clear, however, in the cited studies reporting right hemispheric or bilateral activations in aphasic patients, whether these relate to semantic or to lexical or pre-lexical phonological processing, as these levels of processing were not subtracted away in the critical comparisons. In our normal group study we were able to show that hemispheric lateralization within the temporal lobe depended on the level of auditory word processing, Downloaded by [University of West Florida] at 22:02 09 October 2014 384 R. Zahn et al. i.e. the control tasks used for subtraction. Left lateralization was observed for semantic word processing, whereas word form processing led to bilateral temporal activations (Zahn et al., 2000) in accordance with evidence from clinical studies [reviewed in Kinsbourne (1998)] as well as a previous functional activation study (Mummery et al., 1999a). Consequently right hemispheric activity in patients is not necessarily the result of a left-to-right hemispheric shift as long as it has not been demonstrated to be specifically related to a highly left lateralized function, such as semantic word processing. In a fMRI study, Calvert et al. (2000) actually showed differential lateralization of activations for phonological (to the right) and semantic (to the left) processing after recovery from aphasia. Thalamic activations in RC could be involved in lexical retrieval processes, as suggested by clinical observations of word-finding difficulties after left thalamic stroke (Sodeyama et al., 1995). Increases in the pre-cuneus have been observed in visual imagery tasks (Kosslyn et al., 1995) and in semantic tasks that were likely to engage visual imagery (Démonet et al., 1992). Visual imagery could partially compensate for conceptual-semantic deficits in RC, as all stimuli in the present task denoted concrete entities. Finally, methodological limitations of the present and previous functional activation studies in aphasic stroke patients have to be discussed. Most studies have compared patients with groups or subgroups of normals, and do not report single normal subject data. This might be due to the unsolved problem of how to interpret the considerable interindividual variability in single subject data: functional or anatomical variability (Steinmetz and Seitz, 1991), different cognitive strategies or low statistical power in the single case. In our single subject analysis, we found that although posterior parietal activations were highly significant in the conjunction over subjects, on a single case basis activity was detectable in only eight of 14 single subjects at P ⫽ 0.01. This might be due to the higher type two error in the single case analysis. Therefore, the lack of posterior parietal activation in RC cannot be viewed as abnormal. Our approach of using anatomically non-normalized patient data led to a higher precision in anatomical localization and detection of perilesional activations. It, however, precluded a direct statistical comparison of patient and normal group data as in Price et al. (1998, 1999). Another point of criticism could be that our patients performed below the normal range in the semantic tasks. This could have led to abnormal activation patterns due to abnormal performance that are not comparable with activation patterns in normals (Price and Friston, 1999). Our interpretation of deviant functional activations in RC and HP can thus only concern the pathophysiology of recovery from TSA and not the functional anatomy of semantic processing in normals. In our view, as long as patients’ performances demonstrate that they are specifically engaged in the task and performing significantly above chance level (in our case two standard deviations above, i.e. d’ around 2), one is still measuring a specific response to the task in question and can thus derive valid conclusions. Looking at our patients’ data it is surprising, given their subnormal performance, that their activation patterns are comparable with normal subjects’ activations with respect to important aspects: demonstration of (1) left hemispheric lateralization (RC and HP), (2) posterior inferior and middle pre-frontal activations (RC) and (3) posterior middle/inferior temporal activations (HP) (also detectable in eight of 14 normal subjects). Our main conclusions concerning the organization of the normal semantic system are still valid when only these normal aspects of functional activation patterns in RC and HP are considered. Although our, as well as other, functional activation studies in aphasic patients do have several methodological limitations, a careful interpretation of the data can still lead to valid conclusions, in our view. General discussion The functional activation studies showed posterior inferior parietal, parieto-occipital, parieto-temporal and temporooccipital activations as correlates of semantic processing (Démonet et al., 1992; Vandenberghe et al., 1996; Warburton et al., 1996; Binder et al., 1997). Semantic deficits have been observed in different patients with non-overlapping lesions in the temporolimbic cortex after herpes simplex encephalitis [reviewed in Gainotti et al. (1995)], and after left temporooccipital (Kertesz et al., 1982; Alexander et al., 1989) and parieto-occipital ischaemia (Kertesz et al., 1982). In neurodegenerative dementias, semantic impairment correlated with left inferior temporal (Mummery et al., 1999b) and left parietal (Grossman et al., 1997; Desgranges et al., 1998) and pre-frontal (Desgranges et al., 1998) affection. The anatomical variability of reported lesions associated with semantic impairments could either be explained on the basis of interindividual variability (Steinmetz and Seitz, 1991; Frackowiak, 1997; Berthier, 2001) or attributed to the distributed nature of the system itself (Alexander et al., 1989). Although there is evidence for specific subfunctions of anatomical regions within this system (Gainotti et al., 1995), our study provided evidence for at least limited redundancy, suggesting close relationships between functionally differentiated areas. Conclusions The present study showed that damage to only pre-frontal or to pre-frontal and posterior parietal areas activated in normal subjects during conceptual-semantic word processing led to acute TSA in two consecutive cases. The rapid recovery of language comprehension was associated with left hemispheric activations, in both cases parts of Wernicke’s area. Wernicke’s area, although only activated in one normal subject in this contrast, has been demonstrated as relevant to language comprehension in numerous previous studies. This suggests that a redundancy recovery mechanism rather than a vicari- Recovery of semantic word processing in TSA 385 ation process was in operation in our cases. Redundancy recovery within the left hemisphere supports distributed models of semantic processes and multiple left hemispheric representations of closely related functions. A lesion confined to only part of such a distributed semantic system seems to be partially compensated for by intact areas of the same system with related functions. Further prospective clinical studies are needed to assess the frequency of auditory comprehension deficits in the acute phase after left prefrontal lesions and their recovery. Downloaded by [University of West Florida] at 22:02 09 October 2014 Acknowledgements We thank Professor Dr J. Noth, Dr S. Fellows and two anonymous reviewers for their valuable comments on the manuscript. This study was supported by a grant to MS from the Interdisciplinary Research Centre for the Pathogenesis of Cognitive Disorders at the RWTH Aachen. References Alexander MP, Hiltbrunner B, Fischer RS. Distributed anatomy of transcortical sensory aphasia. Archives of Neurology 1989; 46: 885–92. Berthier ML. Review. Unexpected brain–language relationships in aphasia: Evidence from transcortical sensory aphasia associated with frontal lobe lesions. Aphasiology 2001; 15: 99–130. Binder JR, Frost JA, Hammeke TA, Cox RW, Rao SM, Prieto T. Human brain language areas identified by functional magnetic resonance imaging. Journal of Neuroscience 1997; 17: 353–62. Boatman D, Gordon B, Hart J, Selnes O, Miglioretti D, Lenz F. Transcortical sensory aphasia: revisited and revised. Brain 2000; 123: 1634–42. Buckner RL, Corbetta M, Schatz J, Raichle ME, Petersen SE. Preserved speech abilities and compensation following prefrontal damage. Proceedings of the National Academy of Sciences of the USA 1996; 93: 1249–53. Calvert GA, Brammer MJ, Morris RG, Williams SC, King N, Matthews PM. Using fMRI to study recovery from acquired dysphasia. Brain and Language 2000; 71: 391–9. Cao Y, Vikingstad EM, George KP, Johnson AF, Welch KM. Cortical language activation in stroke patients recovering from aphasia with functional MRI. Stroke 1999; 30: 2331–40. Cappa SF, Perani D, Grassi F, Bressi S, Alberoni M, Franceschi M et al. A PET follow-up study of recovery after stroke in acute aphasics. Brain and Language 1997; 56: 55–67. Demonet JF, Chollet F, Ramsay S, Cardebat D, Nespoulous JL, Wise R et al. The anatomy of phonological and semantic processing in normal subjects. Brain 1992; 115: 1753–68. Desgranges B, Baron JC, de la Sayette V, Petit Taboue MC, Benali K, Landeau B et al. The neural substrates of memory systems impairment in Alzheimer’s disease. A PET study of resting brain glucose utilization. Brain 1998; 121: 611–31. Evans AC, Kamber M, Collins DL, MacDonald D. An MRI-based probabilistic atlas of neuroanatomy. In: Shorvon SD, Fish DK, Andermann F, Bydder GM, Stefan H, editors. Magnetic resonance scanning and epilepsy. New York: Plenum Press, 1994: 263–74. Feeney DM, Baron J-C. Diaschisis. Stroke 1986; 17: 817–30. Finger S, Almli CR. Brain damage and neuroplasticity: mechanisms of recovery or development? Brain Research 1985; 357: 177–86. Frackowiak RS. The cerebral basis of functional recovery. In: Frackowiak RS, Friston KJ, Frith CD, Dolan RJ, Mazziotta JC, editors. Human brain function. San Diego: Academic Press, 1997: 275–300. Friston KJ, Ashburner J, Frith CD. Spatial registration and normalization of images. Human Brain Mapping 1995; 2: 1–25. Gabrieli JD, Poldrack RA, Desmond JE. The role of left prefrontal cortex in language and memory. Proceedings of the National Academy of Sciences of the USA 1998; 95: 906–13. Gainotti G, Silveri MC, Daniele A, Giustolisi L. Neuroanatomical correlates of category-specific semantic disorders: a critical survey. Memory 1995; 3: 247–64. Geschwind N. Disconnexion syndromes in animals and man. Brain 1965; 88: 237–94. Grossman M, Payer F, Onishi K, White-Devine T, Morrison D, D’Esposito M et al. Constraints on the cerebral basis for semantic processing from neuroimaging studies of Alzheimer’s disease. Journal of Neurology, Neurosurgery and Psychiatry 1997; 63: 152–8. Heiss WD, Kessler J, Thiel A, Ghaemi M, Karbe H. Differential capacity of left and right hemispheric areas for compensation of aphasia. Annals of Neurology 1999; 45: 430–8. Hillert D. On processing lexical meanings in aphasia and Alzheimer’s disease: some (re)considerations. Brain and Language 1999; 69: 95–118. Huber W, Poeck K, Willmes K. The Aachen Aphasia Test. Advances in Neurology 1984; 42: 291–303. Karbe H, Thiel A, Weberluxenburger G, Herholz K, Kessler J, Heiss WD. Brain plasticity in poststroke aphasia: What is the contribution of the right hemisphere? Brain and Language 1998; 64: 215–30. Kertesz A, Sheppard A, MacKenzie R. Localization in transcortical sensory aphasia. Archives of Neurology 1982; 39: 475–8. Kinsbourne M. The right hemisphere and recovery from aphasia. In: Stemmer B, Whitaker HA, editors. Handbook of neurolinguistics. San Diego: Academic Press, 1998: 386–94. Kosslyn SM, Thompson WL, Kim IJ, Alpert NM. Topographical representations of mental images in primary visual cortex. Nature 1995; 378: 496–8. Lichtheim L. On aphasia. Brain 1885; 7: 433–84. Mai JK, Assheuer J, Paxinos G. Atlas of the human brain. San Diego: Academic Press, 1998. Marshall JF. Brain function: neural adaptations and recovery from injury. Annual Review of Psychology 1984; 35: 277–308. Mesulam MM. From sensation to cognition. Brain 1998; 121: 1013–52. Miura K, Nakamura Y, Miura F, Yamada I, Takahashi M, Yoshikawa A et al. Functional magnetic resonance imaging to word generation task in a patient with Broca’s aphasia. Journal of Neurology 1999; 246: 939–42. Mohr JP, Pessin MS, Finkelstein S, Funkenstein HH, Duncan GW, Davis KR. Broca aphasia: pathologic and clinical. Neurology 1978; 28: 311–24. Mummery CJ, Ashburner J, Scott SK, Wise RJ. Functional neuroimaging of speech perception in six normal and two aphasic subjects. Journal of the Acoustical Society of America 1999a; 106: 449–57. Mummery CJ, Patterson K, Wise RJ, Vandenbergh R, Price CJ, Hodges JR. Disrupted temporal lobe connections in semantic dementia. Brain 1999b; 122: 61–73. Musso M, Weiller C, Kiebel S, Müller SP, Bülau P, Rijntjes M. Traininginduced brain plasticity in aphasia. Brain 1999; 122: 1781–90. Naeser MA, Helm Estabrooks N, Haas G, Auerbach S, Srinivasan M. Relationship between lesion extent in ‘Wernicke’s Area’ on computed tomographic scan and predicting recovery of comprehension in Wernicke’s aphasia. Archives of Neurology 1987; 44: 73–82. Ohyama M, Senda M, Kitamura S, Ishii K, Mishina M, Terashi A. Role of the nondominant hemisphere and undamaged area during word repetition in poststroke aphasics. A PET activation study. Stroke 1996; 27: 897–903. Phillips CG, Zeki S, Barlow HB. Localization of function in the cerebral cortex: past, present and future. Brain 1984; 107: 327–61. Pietrzyk U, Herholz K, Fink GR, Jacobs A, Mielke R, Slansky I et al. An interactive technique for three-dimensional image registration: Validation for PET, SPECT, MRI and CT brain studies. Journal of Nuclear Medicine 1994; 35: 2011–8. Price CJ. The functional anatomy of word comprehension and production. Trends in Cognitive Sciences 1998; 2: 281–8. Price CJ, Friston KJ. Cognitive conjunctions. A new approach to brain activation experiments. Neuroimage 1997; 5: 261–70. Price CJ, Friston KJ. Scanning patients with tasks they can perform. Human Brain Mapping 1999; 8: 102–8. Price CJ, Howard D, Patterson K, Warburton EA, Friston KJ, Frackowiak RS. A functional neuroimaging description of two deep dyslexic patients. Journal of Cognitive Neuroscience 1998; 10: 303–15. Price CJ, Mummery CJ, Moore CJ, Frakowiak RS, Friston KJ. Delineating necessary and sufficient neural systems with functional imaging studies of neuropsychological patients. Journal of Cognitive Neuroscience 1999; 11: 371–82. Sodeyama N, Tamaki M, Sugishita M. Persistent pure verbal amnesia and transient aphasia after left thalamic infarction. Journal of Neurology 1995; 242: 289–94. Downloaded by [University of West Florida] at 22:02 09 October 2014 386 R. Zahn et al. Springer L, Huber W, Schlenck KJ, Schlenck C. Agrammatism: Deficit or compensation? Consequences for aphasia therapy. Neuropsychological Rehabilitation 2000; 10: 279–309. Steinmetz H, Seitz RJ. Functional anatomy of language processing: neuroimaging and the problem of individual variability. Neuropsychologia 1991; 29: 1149–61. Thompson-Schill SL, D’Esposito M, Aguirre GK, Farah MJ. Role of left inferior prefrontal cortex in retrieval of semantic knowledge: a reevaluation. Proceedings of the National Academy of Sciences of the USA 1997; 94: 14792–7. Thulborn KR, Carpenter PA, Just MA. Plasticity of language-related brain function during recovery from stroke. Stroke 1999; 30: 749–54. Vandenberghe R, Price C, Wise R, Josephs O, Frackowiak RS. Functional anatomy of a common semantic system for words and pictures. Nature 1996; 383: 254–6. Warburton E, Wise RJ, Price CJ, Weiller C, Hadar U, Ramsay S et al. Noun and verb retrieval by normal subjects. Studies with PET. Brain 1996; 119: 159–79. Warburton E, Price CJ, Swinburn K, Wise RJ. Mechanisms of recovery from aphasia: evidence from positron emission tomography studies. Journal of Neurology, Neurosurgery and Psychiatry 1999; 66: 155–61. Weiller C, Isensee C, Rijntjes M, Huber W, Müller S, Bier D et al. Recovery from Wernicke’s aphasia: a positron emission tomographic study. Annals of Neurology 1995; 37: 723–32. Willmes K, Poeck K. To what extent can aphasic syndromes be localized? Brain 1993; 116: 1527–40. Zahn R, Huber W, Drews E, Erberich S, Krings T, Willmes K et al. Hemispheric lateralization at different levels of human auditory word processing: a functional magnetic resonance imaging study. Neuroscience Letters 2000; 287: 195–8. Recovery of semantic word processing in transcortical sensory aphasia: a functional magnetic resonance imaging study R. Zahn, W. Huber, E. Drews, K. Specht, S. Kemeny, W. Reith, K. Willmes and M. Schwarz Abstract In a previous functional magnetic resonance imaging (fMRI) study with normal subjects, we demonstrated regions related to conceptual-semantic word processing around the first frontal sulcus (BA 9) and the posterior parietal lobe (BA 7/40) in agreement with several previous reports. We had the possibility, using the same fMRI paradigm, to study two consecutive cases with left middle cerebral artery (MCA) infarction (RC and HP) and lesions affecting either solely the pre-frontal (HP) or both the pre-frontal and posterior parietal part of the network activated in normal subjects (RC). Both patients showed transcortical sensory aphasia (TSA) on acute assessment. This contradicts classical disconnection accounts of the syndrome stating intact conceptual representations in TSA. Their recovery of language comprehension was associated with activation of a left hemispheric network. Mainly activations of left perilesional pre-frontal regions (RC), left Wernicke’s area (RC and HP) or the left posterior middle and inferior temporal cortex (HP) were demonstrated in the TSA patients. The latter findings suggest that in our cases of TSA functional take-over has occurred in regions with related functions (‘redundancy recovery’) rather than in previously unrelated areas (‘vicarious functioning’). Our data support distributed models of conceptual-semantic word processing and multiple left hemispheric representations of closely related functions. Journal Neurocase 2002; 8: 376–86 Neurocase Reference Number: O271 Primary diagnosis of interest Transcortical sensory aphasia Author’s designation of case RC, HP Key theoretical issue d Recovery of conceptual–semantic word processing Key words: functional magnetic resonance imaging; semantics; pre-frontal cortex; Wernicke’s area; aphasia physiopathology; aphasia transcortical sensory physiopathology; recovery; stroke Scan, EEG and related measures Fluid-attenuated-inversion-recovery-, 3-D MRI, functional MRI Standardized assessment Aachen Aphasia Test Other assessment Neurological examination, Edinburgh Handedness Inventory Lesion location d Left prefrontal (HP, RC) and posterior parietal (RC) Lesion type Ischaemic cardio-embolic stroke Language English