Activation of right fronto-temporal cortex characterizes the ‘living’ category in semantic processing

Cognitive Brain Research 12 (2001) 425–430 www.elsevier.com / locate / bres

Research report

Activation of right fronto-temporal cortex characterizes the ‘living’ category in semantic processing Dirk T. Leube b

a,b ,

*, Michael Erb b , Wolfgang Grodd b , Mathias Bartels a , Tilo T.J. Kircher a

a Department of Psychiatry, University of Tuebingen, Tuebingen, Germany Section Experimental MR of the CNS, Department of Neuroradiology, University of Tuebingen, Hoppe-Seyler-Strasse 3, 72076 Tuebingen, Germany

Accepted 29 May 2001

Abstract It is a vital ability for humans to distinguish between living and non-living objects. Whether the semantic features of these two classes of objects are represented in distinct brain areas, is unknown. In our study, words belonging to the categories ‘living’ and ‘non-living’ were presented visually to twelve right-handed volunteers, while brain activation was measured with event-related fMRI. Subjects had to judge whether the item belonged to one of these categories. Common areas of activation (P,0.05, corrected) during processing of both categories include the inferior occipital gyri bilaterally (BA 17 / 18), left inferior frontal gyrus (BA 44 / 45) and left inferior parietal lobe (BA 40). During processing of ‘living’ minus ‘non-living’ items, signal changes (P,0.05, corrected) were present in the the right inferior frontal (BA 47), middle temporal (BA 21) and fusiform gyrus (BA 19). Our results are in line with findings from patients with a deficit in semantic processing of living things, who specifically suffer from right hemispheric lesions.  2001 Elsevier Science B.V. All rights reserved. Theme: Neural basis of behavior Topic: Cognition Keywords: Semantic category; Living thing; Frontal cortex; Functional imaging

1. Introduction The ability to distinguish between living things and dead objects is a fundamental biological necessity for humans. Children at 36 weeks of age can already separate animate from inanimate objects [3,24]. This ability points to innate neural circuits that are involved in semantic processing of the ‘living–non-living’ categories. The distinction between living and non-living things is basically a categorization task. Neurocognitive approaches hypothesize that semantic categorization is performed by intercorrelating non-categorical properties of an object such as shape, motion, texture, odor or emotions evoked by it. Thus a category (e.g. car, dog, furniture) is defined by a bundle of these properties (e.g. color, shape, odor) [7,40]. *Corresponding author. Department of Psychiatry, University of Tuebingen, Osianderstrasse 24, D-72076 Tuebingen, Germany. Tel.: 149-7071-298-2684; fax: 149-7071-294-141. E-mail address: [email protected] (D.T. Leube).

Whether an object fits into a specific category depends on the specific set of non-categorical properties evoked by it. Categories are represented as patterns of neuronal activation which are distributed over multiple cerebral areas corresponding to various semantic properties [37]. Living things are defined by properties such as biological motion, a certain responsiveness to external stimuli, agency, emotional reaction in the subject her / himself and many others. The more evolutionarily important a category, the more ‘hardwired’ the intercorrelated properties might be in the brain, such as in the case of the ‘living– non-living’ distinction [7]. Categorization requires support from a variety of discrete and specialized brain areas for detecting and evaluating non-categorical properties. For example, occipito-temporal areas are involved in the processing of motion and orbito-frontal regions in emotional content. Impairment in distinguishing specific categories has been described for patients with circumscribed brain lesions (e.g. living things, body parts, manmade objects) [13,15]. For example

0926-6410 / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved. PII: S0926-6410( 01 )00068-4

426

D.T. Leube et al. / Cognitive Brain Research 12 (2001) 425 – 430

in a fMRI study [22] left prefrontal and parietal areas were found to be activated during verbally processing semantic knowledge about body parts. An impairment for processing the ‘living’ category has mainly been described in patients with bilateral hemispheric damage [10,11,18,20,21,36,40]. In contrast, patients with deficits in the ‘non-living’ category show left hemispheric damage [18,36,38,39]. From these brain lesion studies it became evident that the right hemisphere is crucially involved in the categorization of living things [13]. The specific cerebral areas involved are yet unknown. In our study we wanted to investigate the differences in neuronal activity for semantic processing of ‘living vs. non-living’ items using functional magnetic resonance imaging (fMRI). Subjects had to judge whether a visually presented word belonged to the ‘living’ or ‘non-living’ category. We hypothesize that left-sided language areas previously found to be involved in reading (such as left frontal and temporo-parietal regions), are activated for both item types [4,31,32]. Based on findings from lesion studies we expect a right hemispheric activation only for the ‘living’ category [13,36]. More specifically, features such as taste, emotional reactions [5], social learning for positive and negative reinforcers [34] and social interaction [16] are predominantly associated with the ‘living’ category. Lesion studies [16] and results from electrophysiologic experiments [34] associate these features to the frontal lobe. Therefore we expect this area to be predominantly involved during judgment regarding the ‘living’ category.

2. Methods

2.1. Subjects A group of six male and six female right-handed subjects, aged 20–32 (mean 25, S.D. 3.3) years was studied. Subjects were excluded if they had any medical, neurological or psychiatric illness, past or present, or if they were taking medication. After complete description of the study, subjects gave their informed consent.

2.2. Task paradigm Nouns describing a ‘living’ (e.g. horse, child) or a ‘non-living’ category (e.g. train, paper) were presented visually on a screen viewed by the subject via a mirror. The subjects had to press one of two buttons with the right thumb to indicate whether the word belonged to the ‘living’ or ‘non-living’ class. Words of both categories were randomly intermixed. Each word was presented for 4 s followed by a 1 s blank screen. The semantic decision task (ON) alternated with a low level baseline task (OFF) each of them lasting 30 s. A total of five ON and five OFF blocks were presented. During each ON phase, the subjects

were presented with six items, in total 30 words over the whole run. The items were chosen from a list of 500 German nouns. The two groups (living and non-living) were matched for word frequency, length and imagery content [2]. During baseline (OFF), a cross was shown for 4 s followed by a blank screen for 1 s analogous to the word presentation. The subjects had to respond with a button press to each presentation. They were told that they had to remember the words later in another experiment.

2.3. Functional MRI acquisition Imaging was performed on a 1.5 Tesla scanner (Siemens, Vision). Functional images consisted of echoplanar image volumes which were sensitive to Bold contrast (TE 40 ms, TR 2 s). The measurement sequence was modified to allow fast data storing and handling [19]. The volume covered the whole brain with a 64364 matrix and 18 slices (voxel size 33337.5 mm), the slice thickness was 6 mm with a 1.5-mm inter-slice gap. One run consisting of 160 volumes was acquired during the experiment. The first eight acquired volumes were discarded to reach steady state magnetization. A trigger signal from the scanner, the button press of the subject and the onset of the stimuli were registered in a protocol together with the timeline on a separate computer.

2.4. Data analysis For image processing and all statistical analyses SPM99 (Wellcome Department of Cognitive Neurology, London, UK) was used. The functional images of each subject were corrected for motion and realigned by using the first scan of the block as reference. T1 anatomical images were coregistered to the mean of the functional scans and aligned to the SPM T1 template in the Talairach space [35]. The calculated nonlinear transformation was applied to all functional images for spatial normalisation. Finally, the functional images were smoothed with a 12 mm full-width, half-maximum (FWHM) Gaussian filter. Contrasts were calculated by defining SOA (stimulus onset asynchrony) from the protocol as events and by convolving them with the hemodynamic response function (HRF) to specify the appropriate design matrix. Condition and subject effects were estimated according to the general linear model at each voxel in brain space. Global changes of the BOLD signal were removed using proportional scaling. A high pass filter of 120 s and a low pass filter of 4 s were used. Significant hemodynamic changes for each subject and condition were assessed using t-statistics. For the group analysis single subject contrast images were analyzed using a random effects model. The advantage of that approach is that inferences can be drawn not only on the specific group of subjects studied, but also on the whole population they stem from [12]. Contrasts for the ‘living’ and ‘non-living’ conditions were generated and

D.T. Leube et al. / Cognitive Brain Research 12 (2001) 425 – 430

427

Table 1 Regions of significant signal changes as subjects make judgments about ‘living’ items. Activation was measured against a low level baseline and reported if the significance was P,0.05, corrected for multiple comparisons. Talairach coordinates (Tal x, Tal y, Tal z) measured in mm Region

Brodmann area

Side

Tal x

Tal y

Tal z

z value

Inferior occipital gyrus

17 / 18

Left

218

288

21

6.45

Cuneus

18

Right

27

269

17

4.55

Inferior frontal gyrus

47

Left

47 44 / 45 44

Right Left Right

251 233 33 242 39 36

31 20 23 13 19 7

217 29 211 22 27 27

4.73 4.32 4.55 5.62 4.64 4.15

40

Left

236

248

38

6.01

36

245

223

6.01

Inferior parietal lobe Cerebellum

Right

analysed using a simple t-test for each condition and a paired t-test for the subtraction analysis of both conditions. For the conjunction analysis a new contrast image for each subject was generated voxelwise by calculating the minimal t value of each voxel from the two source contrasts. Activations are reported if they exceed P,0.0001 (uncorrected) on the single voxel level and P,0.05 (corrected) on the cluster level. Significance on the cluster level is calculated considering peak activation and extend of the cluster.

3. Results The semantic decision task was performed with a 100% accuracy. In a recognition test 5 min after the end of the session the recognition rate was 90% (S.D. 3%). The activations for ‘living’ items vs. baseline are shown in Table 1, activations for the ‘non-living’ vs. baseline condition in Table 2. The conjunction analysis shows a common activation for both, the ‘living’ and ‘non-living’ condition, in the left inferior frontal gyrus (BA44 / 45), left inferior parietal

lobule (supramarginal gyrus, BA40), left fusiform gyrus (BA 37) and bilateral occipital areas (BA 17 / 18) (Table 3). In the subtraction analysis there was suprathreshold activation only for the ‘living’ minus ‘non-living’ but not for the ‘non-living’ minus ‘living’ subtraction. The ‘living’ minus ‘non-living’ subtraction showed suprathreshold activation in the right inferior frontal gyrus (BA 47,11, Tal x, y, z: 36, 20, 214; z value 5.04) which comprises parts of the ventral part of the prefrontal cortex and parts of the orbitofrontal cortex, the right middle temporal gyrus (BA 21; Tal x, y, z: 51, 27, 217; z value 4.66) and the right fusiform gyrus (BA19; Tal x, y, z: 36, 264, 27; z value 4.33) (Fig. 1).

4. Discussion The main result of our study is a right prefrontal, lateral temporal and fusiform activation when subjects judged words as belonging to the ‘living’ category. There are patients with brain lesions who show a specific semantic deficit either in the ‘living’ or the ‘non-living’ category. A

Table 2 Areas of significant signal changes during judgment of ‘non-living’ items. Activation was measured against a low-level baseline and reported if P,0.05 corrected for multiple comparisons. Talairach coordinates (Tal x, Tal y, Tal z) measured in mm Region

Brodmann area

Side

Tal x

Tal y

Tal z

z value

Inferior occipital gyrus

17 / 18 17 / 18

Left Right

218 21

287 293

21 0

5.48 5.69

Inferior frontal gyrus

44 / 45

Left

254

22

27

5.40

Left

227

220

12

4.51

Left

212 221

230 27

211 222

4.54 4.66

Left

245

235

54

4.34

33

275

237

4.53

Insular cortex Hippocampal formation

28

Inferior parietal lobe

40

Cerebellum

Right

D.T. Leube et al. / Cognitive Brain Research 12 (2001) 425 – 430

428

Table 3 Regions of significant signal changes for the conjunction of both conditions, i.e. judgments about ‘living’ and ‘non-living’ items. Activation is reported if the significance was P,0.05, corrected for multiple comparisons. Talairach coordinates (Tal x, Tal y, Tal z) measured in mm Region

Brodmann area

Side

Tal x,

Tal y

Tal z

z value

Inferior occipital gyrus Inferior occipital gyrus Fusiform gyrus Inferior frontal gyrus Inferior parietal cortex

17 / 18 17 / 18 37 44 / 45 40

Left Right Left Left Left

218 24 239 248 245

287 293 256 10 235

2 0 27 22 54

5.61 5.41 5.91 4.78 4.64

meta-analysis comprising over 50 cases of a ‘living–nonliving’ category impairment [13] showed ‘living-category’ deficits in patients with a bilateral hemispheric damage, whereas only left hemispheric lesions were found in patients with a category-specific impairment for artifacts. Another study [36] systematically investigated 116 subjects with focal unilateral lesions. Lesion overlap for subjects with abnormal retrieval for persons and animals was maximal in the right inferior temporal lobe. Patients with abnormal retrieval of concepts for tools had maximal lesion overlap in the left temporo–parieto–occipital junction. Right inferior frontal damage in cases with a ‘living’ category deficit has been described in several case reports [10,11,18]. Our results of right hemisphere involvement during judgement of items as belonging to the ‘living’ category are in line with the data from lesion studies.

The right frontal, temporal and fusiform areas activated in our study during processing of the ‘living’ minus ‘nonliving’ items point to the existence of multiple object properties which define its categorical classification (OUCH-hypothesis) [7]. No single brain region is responsible for the categorization process, but noncategorical properties of the category member (such as motion, shape, texture, color, odors, emotions) are differentially distributed between the categories of living and non-living things. It refutes the earlier view [40] that objects which are members of the ‘living’ category are predominantly characterized by their perceptual properties. We could demonstrate activation in frontal areas which do not belong to primary perceptual regions. The right frontal activation might be related to the evolutionary aspect of the ‘living–non-living’ distinction.

Fig. 1. Maximum intensity projections (MIPs) of the differential contrast ‘living’ minus ‘non-living’ items in 12 subjects (threshold at P,0.05, corrected for multiple comparisons).

D.T. Leube et al. / Cognitive Brain Research 12 (2001) 425 – 430

It is possible that this discrimination was developed early in phylogenesis and is connected to the emotional response evoked by animals and food intake. Rolls [33,34] showed an involvement of the orbitofrontal cortex in social reinforcement learning. The right orbitofrontal cortex further plays a role in processing emotional content in social contexts. Patients with frontal lobe lesions show aberrant social behaviour [16] and sociopathy was related especially to right frontal damage [5]. The right frontal lobe may thus be related to a specific emotional response, evoked by living objects, which facilitates their categorization. It is striking that brain lesioned patients who show a deficit in the ‘living’ category may also show a semantic deficit in the ‘food’ category [40]. These two categories may share common properties that are processed in the frontal cortex. A food intake disorder with indiscriminate eating can be a clinical feature of frontal lobe dementia [23]. The temporal lobe activation is situated in the superior temporal sulcus (STS). The STS is closely associated with the processing of socially meaningful stimuli concerning gaze, gesture and perception of complex body movements [17,28,30]. The STS is sensitive to stimuli that signal dispositions and intentions of other individuals [14]. These features are essential for the classification of objects as ‘living’. The results of the conjunction analysis showed activation in left inferior frontal gyrus, left fusiform gyrus, left inferior parietal lobe and bilateral inferior occipital lobe during the semantic decision task for both items. This confirms results of earlier PET and fMRI studies which investigated activation during reading or semantic processing [4,31,32]. Neuroimaging studies that investigate ‘living–non-living’ category dissociations [6,8,9,25–27,29] used task designs which required only shallow (e.g. word matching tasks) or feature based (naming and attribute choosing) semantic tasks. The difference between object domains (living vs. non-living) was never drawn to the subjects’ attention, thus remaining implicit. In our study we made domain membership an explicit focus of judgments for subjects. The right hemisphere activations during processing of the ‘living’ category in our study are at variance with results from two early PET studies who found a predominantly left hemispheric activation comparing ‘animate’ vs. ‘inanimate’ stimulus material [9,25]. Both studies examined lexical retrieval in a naming task (i.e. knowing the name of the entity) of animals and tools which were presented as pictures. The same task was later repeated in an fMRI study [8] and there was bilateral involvement of lateral fusiform areas for naming animals in comparison to a more medial bilateral fusiform activation for naming tools. This is in accord with our data as we also found a lateral fusiform activation in the case of living things. Chao et al. [8] also reported that animal stimuli activated a region in the superior temporal sulcus (STS)

429

which was more common in the right hemisphere. In contrast the middle temporal gyrus showed greater activity for naming tools on the left side in their study. It is likely that naming animals and answering questions about their properties is cognitively different from a ‘living–non-living’ domain classification as in our task which focuses much more on features of the objects that are critical for distinguishing among these categories, i.e. features that define the ‘living’ category. In our case we were interested in neuronal activation involved in concept formation rather than lexical retrieval. Other studies found rather conflicting results regarding a right hemisphere or frontal activation by pictures and words from the ‘animate–inanimate’ category. Some [29] find right hemisphere activation in the lingual, superior parietal lobe and inferior temporal gyrus during a word matching task. The relative simple matching to sample task (same–different comparison) may elicit only shallow semantic processing what might explain the absence of frontal activation. Others [26] found right hemisphere activation during the ‘living’ condition including bilateral anterior temporal and right posterior middle temporal cortices, but results have to be viewed cautiously because right hemisphere activation might indicate increased demands on object identification. A study [6] that used word stimuli and a semantic task found frontal and medial temporal activation for the ‘living’ category (BA 10, BA 37) in the right hemisphere. Brain areas activated in our study during the judgment of ‘living’ items may be a correlate of processing the non-categorical properties of the ‘living’ category. The frontal areas together with regions in the temporal lobe (STS) form a network underlying social perception and cognition [1]. Activation of these areas might be the neural basis on which a classification into the ‘living’ category can be performed.

References [1] T. Allison, A. Puce, G. McCarthy, Social perception from visual cues: role of the STS region, Trends Cogn. Sci. 4 (2000) 267–278. [2] I.L. Baschek, J. Bredenkamp, B. Oehrle, W. Wippich, Bestimmung der Bildhaftigkeit (I), Konkretheit (C) und der Bedeutungshaltigkeit (m9) von 800 Substantiven, Zeitschrift Exp. Angewandte Psychol. 24 (1977) 353–396. [3] B.I. Bertenthal, D.R. Proffitt, N.B. Spetner, M.A. Thomas, The development of infant sensitivity to biomechanical motions, Child Dev. 56 (1985) 531–543. [4] J.R. Binder, J.A. Frost, T.A. Hammeke, R.W. Cox, S.M. Rao, T. Prieto, Human brain language areas identified by functional magnetic resonance imaging, J. Neurosci. 17 (1997) 353–362. [5] R.J. Blair, L. Cipolotti, Impaired social response reversal. A case of ‘acquired sociopathy’, Brain 123 (2000) 1122–1141. [6] S.F. Cappa, D. Perani, T. Schnur, M. Tettamanti, F. Fazio, The effects of semantic category and knowledge type on lexical-semantic access: a PET study, Neuroimage 8 (1998) 350–359. [7] A. Caramazza, J.R. Shelton, Domain-specific knowledge systems in

430

[8]

[9]

[10]

[11]

[12] [13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

D.T. Leube et al. / Cognitive Brain Research 12 (2001) 425 – 430 the brain the animate–inanimate distinction, J. Cogn. Neurosci. 10 (1998) 1–34. L.L. Chao, J.V. Haxby, A. Martin, Attribute-based neural substrates in temporal cortex for perceiving and knowing about objects, Nat. Neurosci. 2 (1999) 913–919. H. Damasio, T.J. Grabowski, D. Tranel, R.D. Hichwa, A.R. Damasio, A neural basis for lexical retrieval, Nature 380 (1996) 499–505. M.J. Farah, K.M. Hammond, Z. Mehta, G. Ratcliff, Categoryspecificity and modality-specificity in semantic memory, Neuropsychologia 27 (1989) 193–200. M.J. Farah, P.A. McMullen, M.M. Meyer, Can recognition of living things be selectively impaired?, Neuropsychologia 29 (1991) 185– 193. K.J. Friston, A.P. Holmes, K.J. Worsley, How many subjects constitute a study?, Neuroimage 10 (1999) 1–5. G. Gainotti, What the locus of brain lesion tells us about the nature of the cognitive defect underlying category-specific disorders: a review, Cortex 36 (2000) 539–559. H.L. Gallagher, F. Happe, N. Brunswick, P.C. Fletcher, U. Frith, C.D. Frith, Reading the mind in cartoons and stories: an fMRI study of ‘theory of mind’ in verbal and nonverbal tasks, Neuropsychologia 38 (2000) 11–21. T.M. Gorno, L. Cipolotti, C.J. Price, Category differences in brain activation studies: where do they come from?, Proc. R. Soc. Lond. B. Biolog. Sci. 267 (2000) 1253–1258. J. Grafman, K. Schwab, D. Warden, A. Pridgen, H.R. Brown, A.M. Salazar, Frontal lobe injuries, violence and aggression: a report of the Vietnam head injury study, Neurology 46 (1996) 1231–1238. E. Grossman, M. Donnelly, R. Price, D. Pickens, V. Morgan, G. Neighbor, R. Blake, Brain areas involved in perception of biological motion, J. Cogn. Neurosci. 12 (2000) 711–720. A.E. Hillis, A. Caramazza, Category-specific naming and comprehension impairment: a double dissociation, Brain 114 (1991) 2081–2094. U. Klose, M. Erb, D. Wildgruber, E. Muller, W. Grodd, Improvement of the acquisition of a large amount of MR images on a conventional whole body system, Magn. Reson. Imaging 17 (1999) 471–474. M. Laiacona, R. Barbarotto, E. Capitani, Perceptual and associative knowledge in category specific impairment of semantic memory: a study of two cases, Cortex 29 (1993) 727–740. M. Laiacona, E. Capitani, R. Barbarotto, Semantic category dissociations: a longitudinal study of two cases, Cortex 33 (1997) 441–461. G. Le Clec’H, S. Dehaene, L. Cohen, J. Mehler, E. Dupoux, J.B. Poline, S. Lehericy, P.F. van de Moortele, D. Le Bihan, Distinct cortical areas for names of numbers and body parts independent of language and input modality, Neuroimage 12 (2000) 381–391. M.D. Lezak, in: Neuropsychological Assessment, Oxford University Press, Oxford, 1995, pp. 218–221.

[24] J.M. Mandler, How to build a baby. II. Conceptual primitives, Psychol. Rev. 99 (1992) 587–604. [25] A. Martin, C.L. Wiggs, L.G. Ungerleider, J.V. Haxby, Neural correlates of category-specific knowledge, Nature 379 (1996) 649– 652. [26] C.J. Moore, C.J. Price, A functional neuroimaging study of the variables that generate category-specific object processing differences, Brain 122 (1999) 943–962. [27] C.J. Mummery, K. Patterson, J.R. Hodges, C.J. Price, Functional neuroanatomy of the semantic system: divisible by what?, J. Cogn. Neurosci. 10 (1998) 766–777. [28] H.J. Neville, D. Bavelier, D. Corina, J. Rauschecker, A. Karni, A. Lalwani, A. Braun, V. Clark, P. Jezzard, R. Turner, Cerebral organization for language in deaf and hearing subjects: biological constraints and effects of experience, Proc. Natl. Acad. Sci. USA 95 (1998) 922–929. [29] D. Perani, T. Schnur, M. Tettamanti, T.M. Gorno, S.F. Cappa, F. Fazio, Word and picture matching: a PET study of semantic category effects, Neuropsychologia 37 (1999) 293–306. [30] D.I. Perrett, J.K. Hietanen, M.W. Oram, P.J. Benson, Organization and functions of cells responsive to faces in the temporal cortex, Philos. Trans. R. Soc. Lond. B Biol. Sci. 335 (1992) 23–30. [31] C.J. Price, R.J. Wise, J.D. Watson, K. Patterson, D. Howard, R.S. Frackowiak, Brain activity during reading. The effects of exposure duration and task, Brain 117 (1994) 1255–1269. [32] K.R. Pugh, B.A. Shaywitz, S.E. Shaywitz, R.T. Constable, P. Skudlarski, R.K. Fulbright, R.A. Bronen, D.P. Shankweiler, L. Katz, J.M. Fletcher, J.C. Gore, Cerebral organization of component processes in reading, Brain 119 (1996) 1221–1238. [33] E.T. Rolls, Memory systems in the brain, Annu. Rev. Psychol. 51 (2000) 599–630. [34] E.T. Rolls, The orbitofrontal cortex and reward, Cereb. Cortex 10 (2000) 284–294. [35] J. Talairach, P. Tournoux, Co-planar Stereotaxic Atlas of the Human Brain, Thieme, New York, 1988. [36] D. Tranel, H. Damasio, A.R. Damasio, A neural basis for the retrieval of conceptual knowledge, Neuropsychologia 35 (1997) 1319–1327. [37] L.K. Tyler, H.E. Moss, M.R. Durrant-Peatfield, J.P. Levy, Conceptual structure and the structure of concepts: a distributed account of category-specific deficits, Brain Lang. 75 (2000) 195–231. [38] E.K. Warrington, R. McCarthy, Category specific access dysphasia, Brain 106 (1983) 859–878. [39] E.K. Warrington, R.A. McCarthy, Categories of knowledge. Further fractionations and an attempted integration, Brain 110 (1987) 1273– 1296. [40] E.K. Warrington, T. Shallice, Category specific semantic impairments, Brain 107 (1984) 829–854.