Brain and Language 96 (2006) 69–77 www.elsevier.com/locate/b&l Impairment of syntax and lexical semantics in a patient with bilateral paramedian thalamic infarction Lieve De Witte a, Ineke Wilssens b, Sebastiaan Engelborghs c,d, Peter P. De Deyn c,d, Peter Mariën a,c,d,¤ a Department of Linguistics, Free University of Brussels, Belgium Department of Speech Pathology, Middelheim General Hospital, Antwerp, Belgium c Department of Neurology, Middelheim General Hospital, Antwerp, Belgium d Laboratory of Neurochemistry and Behavior, Institute of Born-Bunge, University of Antwerp, Belgium b Accepted 30 August 2005 Available online 13 October 2005 Abstract Bilateral vascular thalamic lesions are rare. Although a variety of neurobehavioral manifestations have been described, the literature is less documented with regard to accompanying linguistic disturbances. This article presents an in-depth neurolinguistic analysis of the language symptoms of a patient who incurred bilateral paramedian ischemic damage of the thalamus. In the post-acute phase of the stroke, a unique combination of transcortical sensory aphasia with syntactic impairment was found. Because of this atypical semiological association, additional analyses of spontaneous speech were performed. In spite of the typological aYnity with the grammatic characteristic of marked simpliWcation of syntax observed in Broca’s aphasia, only a wordclass speciWc, lexical-semantic deWcit for verbs was objectiWed. The hypothesis that lexical-semantic disturbances in our patient might result from a functional deaVerentiation of both thalami with the frontal lobe is supported by: (1) associated neuropsychological deWcits of frontal origin and (2) frontal-like behavioral disturbances.  2005 Elsevier Inc. All rights reserved. Keywords: Bithalamic stroke; SimpliWcation of syntax; Transcortical sensory aphasia 1. Introduction A variety of neurobehavioral symptoms have been reported in association with bilateral lesions of the thalamus: (1) transient vigilance disorders, (2) hypersomnia, (3) vertical gaze palsy, (4) persistent anterograde amnesia, (5) speech and language disturbances, and (6) apathy (Bewermeyer, Dreesbach, Rackl, Neveling, & Heiss, 1985; Bogousslavsky, Regli, Delaloye-bischof, Assal, & Uske, 1991; Dehaene, 1982; van Domburg, ten Donkelaar, & Notermans, 1996; Fukatsu, Fujii, Yamadori, Nagasawa, & Sakurai, 1997; Guberman & Stuss, 1983; Krolak-Salmon, Croisile, Houzard, Setiey, & Girard-Madoux, 2000; Reilly, Connolly, Stack, Martin, & Hutchinson, 1992). Although speech and language functions have not been systematically * Corresponding author. E-mail address: petermarien@skynet.be (P. Mariën). 0093-934X/$ - see front matter  2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bandl.2005.08.011 investigated, dysarthric and aphasic disturbances have been reported after bithalamic lesions (Lhermitte, 1984). On the articulatory level, the deviant speech characteristics have been denoted as hypokinetic dysarthria (Ackermann, Ziegler, & Petersen, 1993; Benabdeljlil, El alaoui faris, Kissani, Aidi, & Laaouina, 2001; Bewermeyer et al., 1985; Dehaene, 1982; Gentilini, De Renzi, & Crisi, 1987; Gerber & Gudesblatt, 1986; Lazzarino & Nicolai, 1988; Malamut, GraV-Radford, Chawluk, Grossman, & Gur, 1992; Reilly et al., 1992). Reduced spontaneous speech and diminished verbal Xuency have been described at the linguistic level (Bogousslavsky et al., 1991; van Domburg et al., 1996; Engelborghs, Marien, Pickut, Verstraeten, & De Deyn, 2000; Guberman & Stuss, 1983; Krolak-salmon et al., 2000; Krolak-Salmon, Montavont, Hermier, Milliery, & Vighetto, 2002; Malamut et al., 1992; Rousseaux et al., 1986; Schott, Maugiere, Laurent, Serclerat, & Fischer, 1980). In addition, some authors have reported naming 70 L. De Witte et al. / Brain and Language 96 (2006) 69–77 diYculties and mild auditory-verbal comprehension disturbances (Bogousslavsky et al., 1991; van Domburg et al., 1996; Engelborghs et al., 2000; Guberman & Stuss, 1983). We report the Wndings of in-depth neurolinguistic investigations in a patient who incurred transcortical sensory aphasia (TSA) associated with marked simpliWcation of syntax due to a bilateral paramedian thalamic infarction. The combination of Xuent aphasia with syntax simpliWcation is rather unexpected since this grammatical characteristic classically relates to non-Xuent aphasia types. To the best of our knowledge simpliWcation of syntax, characterized in our patient by simple sentences and sentence fragments with a complete absence of embedded clauses, has never been reported in association with transcortical sensory aphasia. 2. Case report A 78-year-old right-handed woman was admitted to our hospital after acute loss of strength in the right arm, diplopia, and speech and language disturbances. Medical antecedents consisted of arterial hypertension and hypercholesterolemia. In addition to Xuent aphasia, the clinical neurological examination on admission revealed vertical gaze palsy and restricted adduction of the left eye resulting in divergent strabism due to left oculomotor nerve paresis. Except for a mild paresis of the right arm, muscle strength and tone were completely normal. The Wnger-tonose test showed dysmetria of the right arm, the knee-toheel test was normal. Tendon reXexes at the right arm and leg were increased. Plantar responses were indiVerent. Sensory and general physical examination revealed no abnormalities. Computerized tomography (CT) scan of the brain disclosed a bilateral paramedian thalamic infarction. Brain magnetic resonance imaging (MRI) demonstrated bilateral involvement of the medial thalamic nucleus (Engelborghs, Marien, Martin, & De Deyn, 1998; Kretschmann & Weinrich, 1992) as well as a small hyperintense area in the left internal capsula. Extension of the left thalamic lesion to the left red nucleus was shown as well (Fig. 1). A diagnosis of acute bithalamic stroke with extension to the mesencephalon was made. Vertebral artery dissection and signiWcant stenoses were excluded by magnetic resonance angiography (MRA). Trans-oesophagal echocardiography was normal. Antiaggregant secondary preventive treatment as well as intensive physical and speech therapy were started. After seven weeks the right hemiparesis had recovered completely. The vertical gaze palsy and oculomotor nerve paresis had only partially improved. The aphasia did not resolve. 2.1. Neuropsychological investigations 2.1.1. Methods Language was assessed according to a time-frame model for the study of aphasia (Alexander, 1989; Mazzochi & Vignolo, 1979). In the post-acute phase of the stroke (lesion phase), extensive formal language investigations were per- formed by means of standardized language tests, among which: the Aachen Aphasia Test (AAT) (Graetz, De Bleser, & Willmes, 1990), the Token Test (De Renzi & Vignolo, 1962), the Boston Naming Test (BNT) (Kaplan, Goodglass, & Weintraub, 1983; Marien, Mampaey, Vervaet, Saerens, & De Deyn, 1998), and a semantic verbal Xuency task (unpublished norms). Four weeks after the stroke (lesion phase) analysis of spontaneous speech and syntactic competence was performed by means of a method developed by Vermeulen, Bastiaanse, and Van wageningen (1989) and the Werkwoorden- en Zinnen Test (WEZT; Bastiaanse, Maas, & Rispens, 2000). Cognitive functions were examined during the lesion phase one and a half month after stroke by means of the Mini Mental State Examination (MMSE) (Folstein, Folstein, & Mc Hugh, 1975), parts of the Alzheimer’s Disease Assessment Scale (ADAS) (Rosen, Mohs, & Davis, 1984), the Coloured Progressive Matrices (Court and Raven, 1983), the Hierarchic Dementia Scale (HDS) (Cole & Dastoor, 1987), the Rey-Osterrieth Figure (Osterrieth, 1944), and the Judgment of Line Orientation test (JLO) (Benton, Hamsher, Varney, & Spreen, 1983). 2.1.2. Neurolinguistic results As shown in Table 1, the AAT results for language comprehension were within the low normal range at both the auditory and written comprehension level. However, as shown by the subtest results, auditory sentence comprehension and written word comprehension were defective, whereas auditory word comprehension and written sentence comprehension were normal. Oral language was Xuent and characterized by word-Wnding diYculties and semantic paraphasias. The spontaneous speech, characterized by simple sentences, sentence fragments with a complete absence of embedded, subordinate clause indicated a marked simpliWcation of syntax (Fig. 2). On the AAT confrontation naming tests a defective global score of 91/120 was found. The patient obtained a normal result for simple nouns. Naming of colors and compound nouns resulted in a deWcient score. Most errors were of the semantic type: semantic paraphasias (e.g., screwdriver D hammer) and semantic neologisms (e.g., cleaner D suckmachine). Although the BNT-score was within the normal range, errors were of aphasic origin: perseverations, circumlocutions (e.g., camel D a seaƒ with two hunches), semantic paraphasias (e.g., octopus D seal; beaver D rat), and semantic neologisms (e.g., protractor D electric measure organ; hammock D lie folder). In a semantic word Xuency task the patient produced a normal number of items. On the AAT sentence construction subtest, in which pictures have to be described depicting complex scenarios, she obtained a defective score. Repetition was normal. Written language, as shown by the AAT results, was normal. Summing-up, the aphasia proWle of the patient corresponds to a taxonomic diagnosis of TSA: language comprehension was disturbed, speech was Xuent and contained a L. De Witte et al. / Brain and Language 96 (2006) 69–77 71 Fig. 1. Brain MRI axial FLAIR slices (A, B, C—from rostral to caudal) performed one month after stroke disclosing bilateral involvement of the medial thalamic nucleus, a small hyperintense area in the left internal capsula and extension of the left thalamic lesion to the left red nucleus. 72 L. De Witte et al. / Brain and Language 96 (2006) 69–77 Table 1 AAT, BNT and word Xuency test results Score Max Mean SD Akense Aphasia Test (AAT) Language comprehension total Auditory word comprehension Auditory sentence comprehension Total auditory comprehension Reading word comprehension Reading sentence comprehension Total reading comprehension 96 27 20 47 21 28 49 120 30 30 60 30 30 60 108.5 26.49 26.79 53.28 28.30 26.91 55.21 10.24 3.30 3.41 6.08 2.29 3.39 4.90 Token Test 13 errors 0 Spontaneous Speech Communicative behaviour Articulation and Prosody Automatisms Semantic structure Phonematic structure Syntactic structure 4 5 4 4 5 2 5 5 5 5 5 5 4.63 4.63 4.59 4.59 4.54 4.41 0.54 0.67 0.65 0.53 0.56 0.55 Imposed Speech Repetition total Phonemes One-syllable words Words of foreign origin Compound words Sentences 147 30 30 29 29 29 150 30 30 30 30 30 144.1 28.91 29.22 28.94 28.45 28.55 8.07 2.09 1.32 2.31 2.22 1.90 Naming total Simple nouns Colors Compound nouns Sentence construction 91 28 23 23 17 120 30 30 30 30 109.3 27.92 27.69 28.04 25.69 8.42 2.90 1.99 2.61 3.72 Written language total Reading aloud Composing on dictation Writing on dictation 80 28 27 25 90 30 30 30 85.52 28.95 28.57 28 7.63 1.93 2.75 3.67 Boston Naming Test 44 60 Verbal Fluency-Semantic Animals Vegetables Clothing Means of transport 35 14 8 9 4 49.1 6.76 43.4 11.76 1 min 1 min 1 min 1 min high incidence of semantic errors but repetition was intact. In addition, spontaneous speech was characterized by marked simpliWcation of syntax. Since the association of TSA with simpliWcation of syntax represents an atypical Wnding, grammatical competence was further investigated. 2.1.3. Additional language analyses 2.1.3.1. Spontaneous language analysis. A sample size of 318 words (Fig. 3) was randomly selected from an extensive corpus of recorded spontaneous speech and compared with the normative data obtained in healthy language users (Vermeulen & Bastiaanse, 1984). Four weeks after the stroke, the corpus was analyzed on both the morphosyntactic and lexical-semantic level. On the morphosyntactic level, mean utterance length (MLU), number of modal verbs (may, must, can, will, ƒ) and copulae (be, become, seem, stay, ƒ), number of subordinate clauses, and the proportion of conjugated verbs were calculated. The number of utterances with conjugated verbs was divided by the number of expressions with one or more verbs. In the lexical-semantic analysis, the number of nouns and lexical verbs as well as the diversity of nouns and lexical verbs represent important variables. To assess lexical diversity, the type token ratio for nouns and lexical verbs was computed. In the type token ratio, which is always calculated on words belonging to the same speciWc word class ( D homogeneous words), the total amount of words belonging to a speciWc word class (noun, lexical verb) is counted Wrst (token). Then the diVerent, diverse words belonging to that speciWc word class are counted ( D type). Each word is deWned as ‘diVerent,’ ‘diverse’ when it occurs for the Wrst time. Morphological forms (e.g., ‘eating’ or ‘dogs’) are not counted as diverse words when already used in another morphological form (‘eat’ or ‘dog,’ respectively). The number of types is then divided by the number of tokens to obtain the type token ratio. Semantic as well as phonological paraphasias were counted as well. At the morfosyntactic level, analysis of the spontaneous speech sample revealed a mean utterance length (MLU) of 4.53 (245/54) words (normal mean: 8.22; range 5.7–10.7). Thirteen copulae and modal verbs were used (normal range: 4–18). Only one subordinate clause was produced (normal range: 3–15). The ratio of conjugated verbs was 0.63 (23/41) (normal range 0.83–1). At the lexical-semantic level the corpus only contained 34 nouns. The type token ratio for nouns was 0.65 (22/34) (normal mean: 0.75; range 0.54–0.88). Thirty-Wve lexical verbs were produced. The type token ratio for lexical verbs was 0.48 (17/34) (normal mean 0.74; range 0.63–0.86). Quantitative analysis of a randomly selected spontaneous speech sample thus showed no marked diVerences between the number of nouns and the number of verbs. The Fig. 2. Structured conversation (literal English translation). L. De Witte et al. / Brain and Language 96 (2006) 69–77 73 Fig. 3. Spontaneous speech sample (literal English translation). diversity in the use of nouns was normal but clearly defective for lexical verbs. Mean utterance length and number of subordinate clauses were deWcient. 2.1.3.2. Grammatical analysis. Comprehension and production of verbs and sentences were investigated by means of the WEZT (Table 2). This test consists of several tasks which allow identiWcation and description of grammatical deWcits. On the receptive level the test is composed of three subtests including verb comprehension, sentence comprehension, and grammaticality judgment; on the expressive level the test consists of seven subtests: action naming, sentence completion with inWnitives, sentence completion with conjugated verbs, oral sentence construction, sentence anagrams, and interrogative sentence anagrams. On the receptive level the ‘verb comprehension task’ examines the ability to comprehend single verbs. In this task the subject is requested to depict one picture from an array of four that matches the verb orally presented by the examiner. The three distractors represent: (1) an object that shares a close semantic relationship with the target verb, (2) an action that is related to the target action and, (3) an object that is related to the action distractor. The 60 items are controlled for frequency, transitivity, and name relatedness with a noun (e.g., ‘to hammer,’ related to ‘hammer’ instead of ‘to knock’). All these features are indicated on Table 2. A deWcient result was found for this subtest (54/ 60). The majority of errors resulted from lexical-semantic problems. The errors consisted of selection of the semantic distractor of the target verb (n D 3) (e.g., target: ‘to squeeze,’ subject’s answer: ‘to sift’ D a related action distractor with no name relationship), selection of the distractor noun that closely relates to the target verb (n D 2) (e.g., target: ‘to grate,’ subject’s answer: ‘a grater’ D a distractor with name relationship), and selection of the distractor noun semantically related to the action distractor (n D 1) (e.g., target: ‘to conjure,’ subject’s answer: ‘a magician’ D a distractor with no name relationship to the target word). The ‘sentence comprehension task’ contains stimulus sentences with moved constituents as in passives (theme, verb, agent) and sentences in which the order of the thematic roles is canonical as in actives and subject-cleft constructions (agent, verb, theme). These sentence types allow the examiner to investigate whether the order of the thematic roles has an impact on sentence comprehension. The procedure used is analogous to the ‘verb comprehension task’: four pictures are presented to the subject while the examiner reads a sentence aloud. The subject is asked to single out the picture that matches the sentence. The distractors consist of: a picture in which the action and the participants are the same as the target picture but in which the thematic roles are reversed, a picture in which a diVerent action is depicted and a picture in which both the thematic roles are reversed and in which the action is diVerent. The result on this subtest was also deWcient (34/40). Three errors consisted of role exchanges (e.g., target: ‘the grandmother photographes the child,’ subject’s answer: ‘the child photographes his grandmother’) and another three errors due to selection of a semantic related action (e.g., target: ‘mother pulls the child,’ subject’s answer: ‘mother hits the child). The ‘grammaticality judgment task’ investigates whether the patient is able to parse a sentence. Only irreversible sentences are tested as they can 74 L. De Witte et al. / Brain and Language 96 (2006) 69–77 Table 2 WEZT lesion phase data WEZT Score Max M Verb comprehension Transitive Name relationship High frequent Intransitive No name relationship Low frequent 54 28 18 26 26 36 28 60 31 20 28 29 40 32 59.27 1.13 SD ASD Cut-oV 4.66 56 Grammaticality judgment Grammatical sentences Non-grammatical sentences 49 24 25 50 25 25 49.70 0.56 1.25 48 Sentence comprehension Verb-agent-theme Agent-theme-verb Agent-verb-theme Theme-agent-verb 34 9 8 8 9 40 10 10 10 10 39.77 0.42 13.7 Action naming Transitive Intransitive Name relationship No name relationship High frequent Low frequent 33 17 16 17 16 18 14 40 20 20 20 20 20 20 38.77 1.73 3.33 32 Filling in inWnitive 8 10 9.57 0.50 3.14 9 Filling in conjugated verbs 9 10 9.65 0.48 1.35 9 Sentence construction Transitive Intransitive Reversible Not reversible 8 4 4 2 2 20 10 10 5 5 16.35 2.95 2.83 9 Sentence anagrams ¡ picture Active Passive 19 9 10 20 10 10 20 0 Sentence anagrams + picture 20 20 20 0 wh-question anagrams + picture 10 Arguments Who/What 5 Adjuncts 5 20 10 10 19.75 0.78 12.5 39 20 0 20 18 be judged correctly without the need to map thematic roles onto the arguments if one has access to the grammatical structure. This subtest yielded a normal result (49/50). In the ‘action naming task’ on the productive level, the patient is asked to name in one word what is happening on the picture shown to him. Forty items are presented that represent agentive verbs. The inXuence of several linguistic parameters (frequency, transitivity, and name relatedness with a noun) can also be determined. The patient obtained a defective score on this subtest (33/40). Most errors consisted of semantic paraphasias (e.g., target: ‘to play at draughts,’ subject’s answer: ‘to play chess’ D an error against an intransitive, low-frequent verb with no name relationship; target: ‘to crawl,’ subject’s answer: ‘to search’ D an error against a transitive, high-frequent verb with no name relationship) and the splitting of the target verb in a semantic related noun and verb (e.g., target: ‘to fence,’ subject’s answer: ‘to play with a sword’ D an error against an intransitive, low-frequent verb with no name relationship). In this task the parameter ‘frequency’ rather than ‘transitivity’ and ‘name relatedness with a noun’ seemed to determine the patient’s performance. She obtained 14/20 for low-frequent verbs and 18/20 for highfrequent verbs. The ‘sentence completion task with inWnitives or conjugated verbs’ comprises two sets of sentences: 10 sentences that require completion with an inWnitive verb and 10 sentences that require a Wnite verb. On the ‘sentence completion task with inWnitives,’ our patient obtained a defective score (8/10). Only semantic errors occurred. In the ‘sentence construction task,’ the patient is instructed to describe a picture in one sentence. The 20 items that build up this test are controlled for three factors: frequency (all verbs are of high frequency), transitivity (10 transitive and 10 intransitive verbs), and reversibility of thematic roles (5 reversible and 5 irreversible sentences). On this subtest a score of 8/20 was found. Expressive agrammatic features (use of inWnite verbs, omission of function words, stereotype sentence construction) and lexical-semantic deWcits disrupted performance at the ‘sentence construction task’ (e.g., woman swim instead of the child swims, father falls instead of the baby crawls) (Fig. 4). ‘The sentence anagram task’ consists of three subparts. In the ‘sentence anagram task with pictures’ the subject is presented a picture and three cards, each of which has a sentence constituent printed on it. The subject is asked to sort these cards into a sentence that matches the picture. This subtest contains 20 items representing 10 reversible and 10 non-reversible sentences. The ‘sentence anagram task without pictures’ contains 20 irreversible sentences of which 10 are active and 10 are passive. In this task, three cards with constituents have to be arranged into a sentence. The ‘interrogative sentence anagram task’ includes 20 items: Wve items for each wh-question word (who, what, where and when). The procedure is similar to the sentence anagram task (Bastiaanse, Edwards, Maas, & Rispens, 2003). Only on the interrogative sentence anagram task the patient obtained a pathological score (10/20) as a result of role exchanges and incorrect word order. No diVerence was found between arguments (5/10) and adjuncts (5/10) (Fig. 5). Summing-up, formal linguistic analysis of expressive and receptive syntax disclosed: (1) a lexical-semantic impairment at the level of the category of verbs, (2) a discrete mapping problem, and (3) a selective disturbance for interrogative sentence constructions. 2.1.4. Neuropsychological test results One and a half month post-onset, cognitive functions were formally investigated. Manual preference was assessed by means of the Edinburgh Handedness Inventory (OldWeld, 1971) which revealed a laterality quotient of +100. Orientation, gnosis, praxis, problem solving, arithmetics, memory, and spatial intelligence were normal. On the Mini Mental State Examination the patient obtained a normal score of 26/30 (Crum, Anthony, Bassett, & Folstein, 1993). She obtained percentile 75 on the Raven Coloured Progressive L. De Witte et al. / Brain and Language 96 (2006) 69–77 75 Fig. 4. Sentence construction (WEZT) (literal English translation). Fig. 5. Wh-interrogative sentence constructions. Matrices and normal results were found for ‘word recall,’ ‘commands,’ ‘naming Wngers and objects,’ ‘constructional praxis,’ ‘ideational praxis,’ ‘orientation,’ and ‘word recognition’ of the Alzheimer’s Disease Assessment Scale (Zec, Landreth, Vicari, Feldman, & Belman, 1992a, 1992b). No indications for a bucco-labio-lingual, dressing, visuoconstructive, ideational or ideomotor apraxia were found. Visuo-spatial processing as well as visual, tactile, and auditory recognition were normal. Normal results were found on the memory items of the Hierarchic Dementia Scale: ‘registration,’ ‘recent memory’ en ‘remote memory.’ Frontal dysfunctions were objectiWed on both the behavioral and psychometric level. As evidenced by a severely deWcient score on the Stroop Color–Word test (