Are regular and irregular verbs dissociated in non-ﬂuent ...

Brain Research Bulletin 74 (2007) 1–13

Review

Are regular and irregular verbs dissociated in non-fluent aphasia?A meta-analysis

Yasmeen Faroqi-Shah ∗Department of Hearing and Speech Sciences, University of Maryland, 0100, Lefrak Hall,

College Park, MD 20742, United States

Received 10 February 2007; received in revised form 7 June 2007; accepted 12 June 2007Available online 5 July 2007

Abstract

The cognitive mechanisms and neuroantomical substrates used by the brain to effortlessly generate morphologically complex words(write + ing → writing) are little understood. The left inferior frontal gyrus (LIFG, including Broca’s area) is often implicated as being involved,although its specific role is unclear. Data from brain damaged individuals, particularly those with Broca’s aphasia, are often used as evidence tosupport or refute various theoretical perspectives. Typically, performance on two types of morphologically complex verbs, regulars (walk-walked,slip-slipped) and irregulars (sing-sang, sleep-slept) is contrasted for evidence of single or double dissociations. The question of how Broca’s aphasicindividuals dissociate in their production of inflectional morphology is important to our understanding of how the brain is organized to computemorphologically complex words. This article is a synthesis of research studies investigating the production of morphologically complex regularand irregular verbs in individuals with Broca’s aphasia. The question being asked is if there is a robust and consistent dissociation, and if thisdissociation can be tied to lesions of the left frontal lobe. This meta-analysis of 75 patients failed to show a single consistent dissociation pattern andover half the datasets had no significant difference between regulars and irregulars. There was also no relationship of any performance pattern tofrontal lobe lesions, highlighting the difficulty of identifying any single neuroanatomical lesion for regular–irregular verb production deficits. Theimplications for various theoretical and neuroanatomical hypotheses are discussed. The role of neuropsychological dissociations in constraininghypothesis of normal neuroanatomical organization is evaluated.© 2007 Elsevier Inc. All rights reserved.

Keywords: Morphology; Broca’s aphasia; Language production; Sentence completion; Repetition; Verb

Contents

1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.1. Neuroanatomical predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2. Neuropsychological dissociations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.3. Purpose of this meta-analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2. Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.1. Coding and data analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.1. Sentence production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.2. Repetition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.3. Lesion information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.4. Inter-test consistency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94.1. Is there a dissociation between regular and irregular verbs? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94.2. Are specific lesions sites associated with specific dissociation patterns? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

∗ Tel.: +1 301 405 4229; fax: +1 301 314 2023.E-mail address: [email protected].

0361-9230/$ – see front matter © 2007 Elsevier Inc. All rights reserved.doi:10.1016/j.brainresbull.2007.06.007

mailto:[email protected]

dx.doi.org/10.1016/j.brainresbull.2007.06.007

2 Y. Faroqi-Shah / Brain Research Bulletin 74 (2007) 1–13

4.3. Can dissociations be manipulated by experimental variables? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105. Conclusions and future research considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Conflict of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

. . . . .

1

oidtsoiwls[sivbh(fbrtb

agqlmpwtlRdh

1

pucltcn

iatpeifwtnnulars and irregulars [49,54]. Hence, neuroimaging evidencefor the declarative/procedural model has been inconclusive[5].

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. Introduction

Few questions about the nature and neural organizationf the human language faculty have generated the kind ofntense debate that word inflections have in the past twoecades. The controversy over regular and irregular word inflec-ions, often called the “English past tense debate” [46], haspanned cognitive neuroscience, linguistics, cognitive psychol-gy, neuropsychology, language development, and artificialntelligence alike. Many languages have two types of inflectedords (although this dichotomy may be simplistic [9]): regu-

ar inflections, such as past tense -ed in English (walk-walked,lip-slipped) and the participle -t in German (tanzen-getanztdance] [45]) and irregular inflections with non-default changes,uch as sing-sang, sleep-slept in English, or the participle -nn German (trinken-getrunken [drink]). Differing opinions pre-ail regarding the nature of linguistic and cognitive differencesetween regular and irregular words. Some believe that compre-ending and producing regular inflections involves affixationwalk + ed → walked), while irregular inflections are retrievedrom memory (dual mechanism account) [12,41,47,59]. Otherselieve that all inflections are retrieved from memory (singleoute, connectionistic accounts) [10,31,48], or are affixed (dis-ributed morphology) [26,54], or that there is a competitionetween both memory and affixation routes [2].

Linguistic and neurocognitive distinctions between regularsnd irregulars have assumed enormous importance among lan-uage researchers because these succinctly embody a broaderuestion about human language: whether we have two distinctanguage modules, a mental grammar and a mental lexicon. Theental grammar is a system of combinatorial rules that can beroductively applied to generate sentences as well as complexords. Regular inflections are considered to be archetypes of

hese combinatorial rules. The mental lexicon is essentially aisting of all known words, including irregular inflections [46].ecently, the question is not so much whether the two wordsiffer in psycholinguistic aspects, but whether these differencesave neuroanatomical validity.

.1. Neuroanatomical predictions

The Declarative/Procedural hypothesis first made distinctredictions about the neural correlates of regular and irreg-lar words [59]. Based on the assumption that grammaticalombinatorial rules operate for the affixation of regulars, the
eft frontal cortex, particularly Broca’s area, was implicated inhe processing and production of regular words. Basal gangliaircuits were also identified in this fronto-striatal proceduraletwork. In contrast, irregular verbs were assumed to be housed
Fpgct

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

n the mental lexicon and retrieved from declarative memory,nd linked to the left temporo-parietal cortex (Fig. 1a). Func-ional neuroimaging studies have tested these neuroanatomicalredictions using a variety of techniques, such as positronmission tomography, functional magnetic resonance imag-ng (FMRI), and magnetoencephalography. Some studies haveound supportive evidence for the declarative/procedural modelith greater left frontal activation for regulars and greater left

emporal activity for irregulars [17,30,44,58]. However, othereuroimaging studies have found the reverse pattern [28], oro remarkable differences in the neural processing of reg-

ig. 1. Predictions of the various neuroanatomical models: (a) declarative-rocedural model [59]; (b) phonology-semantics model [57]; (c) two-stage verbeneration model [5], not shown are: left insula and hippocampus, also impli-ated for regular verbs and right DLPFC, also implicated for irregular verbs. Seeext (Section 1.1.) for details.

search

oamiatirnfpcci(ratotriurra

vittvhtTvcatsg(

lpdfint[ttThr

lol

1

ds[asc([aifcigibrsahufv

noBcwbnnsuald[

sctregular–irregular verb processing imply normal versus impairedperformance on irregular and regular verbs, respectively, in indi-viduals with frontal lobe damage [41,59]. However, operational

Y. Faroqi-Shah / Brain Re

A more recent neuroanatomical hypothesis was proposedn the basis of auditory input processing deficits of regularsnd irregulars [56,57] (hereafter called the semantics-phonologyodel). Examples of input processing tasks include deciding

f auditorily presented stimuli are real words (pless, vaim)nd same-different judgments with word pairs (sprayed-spray,rade-tray, taught-teach). The dorsal–ventral speech process-ng framework [27] was used to propose that the presence ofegular inflectional affixes engages the dorsal (fronto-temporal)etwork to greater extent than irregular inflections [56]. Thisronto-temporal network, which includes bilateral superior tem-oral gyri, left inferior frontal gyrus (LIFG), and left anterioringulate, is implicated for “morphophonological parsing pro-esses that segment morphologically complex spoken formsnto stems and affixes” [56, p. 8380] (sprayed → spray + ed)Fig. 1b). Explicit statements about the neuroanatomical cor-elates of irregulars are not made in earlier versions [56,57],lthough a recent version [42] allocates left posterior superioremporal gyrus and middle temporal gyrus for semantic accessf irregular verbs and monomorphemic words. In other words,his model claims that phonological demands are greater foregulars while semantic access is more crucial for processingrregular verbs and monomorphemic words. An FMRI studysing same-different word pair judgments by the same authorsevealed that regulars and irregulars activated the same left supe-ior temporal–frontal network and only varied in the extent ofctivation (greater activity for regulars) [57].

Another recent data-driven proposal is based on FMRI acti-ation patterns for covert generation of Spanish regular andrregular verbs [5] (referred to as the two-stage verb genera-ion model hereafter). Both regular and irregular verbs activatedhe left opercular LIFG, regular verbs showed increased acti-ation of left anterior superior temporal gyrus-insula andippocampus, while irregular verbs showed increased activa-ion bilaterally in the dorsolateral–prefrontal cortex (DLPFC).he authors proposed a two-stage model for the generation oferb inflections, the first stage is language and verb type spe-ific, and recruits the DLPFC for lexical access of irregulars,nd grammatical–phonological network of left anterior superioremporal gyrus, insula and hippocampus for regular verbs. Theecond step is common to all verb types, involves retrieval oframmatical information, such as tense and recruits the LIFGFig. 1c).

To summarize, the declarative/procedural model proposeseft fronto-striatal involvement for all regulars, and left temporo-arietal involvement for all irregulars, irrespective of taskemands [59]. The semantics-phonology model claims greaterronto-temporal (dorsal) activity for regulars compared torregulars, but does not make any specific statements abouteuroanatomical correlates for irregulars, or for productionasks [56,57]. Finally, the two-stage verb generation model5] suggests that LIFG activity is common to both verbypes, DLPFC is increasingly activated for irregulars and a
emporal-insular-hippocampal network is used for regulars.hese predictions are illustrated in Fig. 1. Neuroimaging studiesave produced inconsistent results and are hence inconclusiveegarding the neuroanatomical correlates of regulars and irregu-
d

c

Bulletin 74 (2007) 1–13 3

ars [16,17,28,30,49,54,56]. In the next section, another sourcef evidence for neural differences between regulars and irregu-ars is evaluated.

.2. Neuropsychological dissociations

A widely cited substantiation for the neuroanatomical pre-ictions of the Declarative/Procedural hypothesis comes fromingle and double dissociations in brain damaged populations41,59]. Although a variety of conditions that predominantlyffect left frontal–striatal structures [Broca’s aphasia, Parkin-on’s disease, Huntington’s disease] have been compared withonditions that predominantly affect posterior language areasanomic aphasia, Wernicke’s aphasia, Alzheimer’s dementia)58,29,60], Broca’s aphasia has received particular attentions the poster-child in the debate between regular and irregularnflections [8,21,41,43]. Broca’s aphasia typically results fromocal brain damage to left hemisphere frontal regions [15] and isharacterized by reduced speech rate, inaccurate grammar, andncorrect usage of verb inflections [50]. Using data from sin-le case studies, it was originally proposed that Broca’s aphasicndividuals, owing to frontal lobe damage, show dissociationetween regulars and irregulars, with worse performance onegulars [41,59]. However, an accumulating body of evidenceuggests that the relationship between regular verb impairmentnd Broca’s aphasia is not straightforward because some studiesave found the reverse pattern of worse performance on irreg-lars compared to regulars [8,6,45,29,43], while others haveound no difference between regulars and irregulars [8,15], or aariety of patterns [22].

Task differences, language differences, and variations in theumber of patients tested are only a few factors that havebscured the relation between regular verb production androca’s aphasia. For example, it can be questioned if disso-iation patterns from cognitively diverse tasks, such as singleord reading, lexical priming, and sentence completion, cane pooled together. Is it reasonable to assume that the putativeeural differences between regulars and irregulars will surpasseuroanatomical differences from inherently dissimilar tasks,uch as reading and sentence completion? Further, unlike irreg-lars in English (swam, kept, etc.), irregulars in languages, suchs Spanish and German contain non-default (albeit quasiregu-ar) affixes, and it is unclear if the brain makes the same binaryistinction between regulars and irregulars in such languages45].

In addition to the extensive methodological diversity acrosstudies, there are conceptual and statistical concerns thatast doubt on the robustness of regular–irregular dissocia-ions in aphasic individuals [7,13].1 Neuroanatomical models of

efinitions of normal and impaired performance or what con-

1 These conceptual and statistical limitations apply to other areas of neuropsy-hological research as well [36].

4 search

stcttuMutbtciIswpia

1

frtftaTiqrdiioiimufr

2

ittfprstdTcd

itrwbtsd

2

tompctIfat

tsuctp

3

fodssiT

3

f(stuinTaap

Other variables. Forty-nine (out of total 78) data sets camefrom English speakers (33 unique patients). Interestingly, 9 out


titutes a dissociation are not provided in most studies, a pointhat has been made about most research on neuropsychologi-al dissociations [14]. It has also been emphasized elsewherehat reliable inferences about normal neurocognitive architec-ure can be made from neuropsychological dissociations onlypon statistical comparisons with a normal control group [14].ost studies of regular–irregular dissociations in aphasia have

sed intra-individual comparisons, without reference to a con-rol group. These so called trend dissociations are considered toe weakest form of neuropsychological evidence [52,38]. Fur-her, it has been demonstrated that tasks that are likely to produceeiling effects in normal controls (100% accuracy) significantlynflate the occurrence of dissociations (high false positives-Type

error] in brain damaged individuals [38]. Finally, conclu-ions about neuroanatomical localization of function are limitedhen patients with similar lesions exhibit reverse dissociationatterns. In other words, neuropsychological dissociations arenformative only if we can clearly demonstrate that the dissoci-tion exists.

.3. Purpose of this meta-analysis

In the presence of mixed evidence and methodological con-ounds, it is unclear if a single dissociation pattern betweenegular and irregular verbs exists in Broca’s aphasia. This ques-ion is important because different neural regions are identifiedor regular–irregular processing by the declarative/procedural,wo-stage verb generation, and auditory input models (Fig. 1),nd it is unclear if existing data favor any one of these models.his scenario warrants a systematic synthesis of research find-

ngs to ascertain the weight of neuropsychological evidence. Theuestion being evaluated in this meta-analysis is not whetheregulars and irregulars can dissociate, rather if they consistentlyissociate in a single direction (regulars worse than irregulars)n patients with Broca’s aphasia, and whether this dissociations consistently associated with left frontal lobe lesions. A sec-ndary purpose was to explore variables that are likely to resultn dissociations. Such a synthesis is likely to yield fruitful resultsf patient performance is compared across similar tasks. Since a

ajority of studies have used production tasks to compare reg-lar and irregular verbs in Broca’s aphasia, this meta-analysisocuses on two production tasks: sentence production and wordepetition.

. Materials and methods

Published articles or abstracts that compared the production of regular andrregular verbs in English peer-reviewed journals were identified by searcheshrough electronic databases. The electronic databases included Science Cita-ion Index, Medline (Pubmed), PsycInfo, and CINAHL. Combinations of theollowing keywords were used to conduct the search: aphasia, verbs, languageroduction, morphology, dissociation, regular, and irregular. The search wasestricted to research studies published between 1980 and October 2006. Theearch resulted in a total of 56 reports. The search was further narrowed down
o 25 articles after reading the abstracts for relevance. Reports that duplicatedata, such as conference presentations and full articles, were included only once.he entire text of these final 25 articles was read and the following inclusionaryriteria were used: (1) the study reported original data obtained from patient(s)escribed as non-fluent, agrammatic or Broca’s aphasic by the author(s); (2)
o

u

Bulletin 74 (2007) 1–13

ndividual subject data were reported; (3) regular and irregular verb produc-ion was tested and the scores of regular and irregular verbs were separatelyeported2; (4) either sentence completion, and/or single word repetition tasksere used. Fifteen studies met these criteria. Of these, three studies reportedoth sentence production and word repetition data, 10 studies reported only sen-ence production data and two had only single word repetition. Multiple dataets from individual patients were included in the meta-analysis only if eachata set was original to the study.

.1. Coding and data analyses

All studies were coded for a variety of variables including, language ofesting, the number of stimuli used, raw scores, proportion accuracy, descriptionf aphasia profile, description of lesion information, phonological and frequencyatching of stimuli, and the overall conclusions of the authors. For sentence

roduction studies, three task variations have been used, these were separatelyoded as A, B, and C. Task A referred to a fill in the blank task where the presentense verb was provided in a preceding context (Everyday I rob a bank. Yesterday

), Task B corresponded to fill in the blank that required selectionrom multiple options (Yesterday I a bank: rob, robbing, robbed),nd Task C elicited the sentence in response to a question (The man is stealinghe silverware. What did he do yesterday?).

For studies where raw data were not reported, these were computed fromhe relevant figures or percentage scores. The authors of one study provided rawcores via personal communication [8]. Raw scores were statistically comparedsing the Fisher-exact test since this test considers the number of stimuli inalculation of p-values [40,51]. The one-tailed p-values thus derived were usedo code if performance of regulars was worse than irregulars since this was theredicted direction of dissociation for Broca’s aphasia [41,59].

. Results

A total of 110 different datasets were obtained across all tasksrom 75 patients. Nineteen individuals contributed more thanne dataset and 16 patients contributed to both sentence pro-uction and word repetition data. A complete list of includedtudies and individual subject data are provided in Table 1 forentence completion and Table 2 for word repetition. The find-ngs of the two tasks are described in separate sections below.his is followed by the results of the lesion analysis.

.1. Sentence production

A total of 78 separate patient data sets, representing 66 dif-erent aphasic patients were obtained for sentence productionTable 1). The number of data sets in which regular verbs wereignificantly lower in accuracy compared to irregular verbs (one-ailed p < 0.05) is 10 (9 unique patients). In 26 data sets (23nique patients), regular verbs were significantly better thanrregulars in accuracy. In the remaining 42 data sets, there wereo significant differences between regular and irregular verbs.hese proportions are summarized in Fig. 2. Overall, there washigh positive correlation between the accuracies of regulars

nd irregulars, when computed across all 78 datasets (rs = 0.64,< 000).

f the 10 patients who had poorer accuracy for regulars were

2 Authors definitions of and decisions about what constitutes “regularity” weresed for the purpose of this analysis.

Y.Faroqi-Shah/B

rainR

esearchB

ulletin74

(2007)1–13

5Table 1Studies of sentence production sorted by language and chronological order

Reference Language Task/stimuli Patient Lesion # Stimuli Accuracy p One-tail R < I? p < 0.05

R I R I

Ullman et al., p. 271 (Table 3) [59] English B FCL L frontal, insula, white matter 20 20 0.20 0.69 0.00 Yes

Bird et al., p. 507 (Fig. 1), p. 509(Fig. 2) [8]

English B F

BB L 44 44 0.00 0.08 0.06 NoIJ L 44 44 0.07 0.23 0.03 YesIB L frontal temporal parietal 44 44 0.09 0.14 0.37 NoDE L mca 44 44 0.20 0.43 0.02 YesAB L mca 44 44 0.24 0.36 0.18 NoMB L frontoparietal, R frontal 44 44 0.18 0.39 0.03 YesRT L 44 44 0.07 0.34 0.00 YesVC L frontoparietal 44 44 0.20 0.36 0.08 NoGN L mca incl. Broca’s, occipital 44 44 0.28 0.51 0.02 YesJL L frontal temporal parietal 44 44 0.69 0.84 0.07 No

English B P F BB 34 34 0.18 0.12 0.85 NoIJ 34 34 0.06 0.06 0.69 NoIB 34 34 0.26 0.09 0.98 NoDE 34 34 0.21 0.26 0.39 NoAB 34 34 0.24 0.18 0.82 NoMB 34 34 0.24 0.26 0.53 NoRT 34 34 0.09 0.24 0.09 NoGN 34 34 0.59 0.05 1.00 NoJL 34 34 0.79 0.62 0.97 No

Shapiro and Caramazza, p. 1193 [53] EnglishB RC L frontal incl. Broca’s, insula,

internal capsule, putamen112 154 0.43 0.18 1.00 No

B F RC 26 26 0.46 0.08 1.00 No

Lambon Ralph et al., p. 110 (Table 4)[36]

English B P F

GD L 44 44 0.67 0.75 0.24 NoPG L mca, parietal 44 44 0.54 0.29 1.09 NoDC L parietal 44 44 0.83 0.75 0.91 NoJS L 44 44 0.75 0.54 0.99 NoDM L subarachanoid hemorrhage 44 44 0.96 0.67 1.00 NoAB 44 44 0.25 0.33 0.24 NoGN 44 44 0.33 0.63 0.01 YesJL 44 44 0.71 0.83 0.10 No

Ullman et al., p. 202 (Table 5) [60] English B (F) RBA L Broca’s, temporal isthmus 20 16 0.20 0.25 0.51 NoDruks p. 1006 [19] English B F MC L 36 36 0.42 0.36 0.77 No

Fix and Thompson, p. 167 (Table 1)*

[22]English B F

P1/B6 21 39 0.43 0.49 0.44 NoP2 21 39 0.71 0.05 1.00 NoP3/B8 21 39 0.71 0.31 1.00 NoP5 21 39 0.29 0.62 0.02 YesP6 21 39 0.60 0.16 1.00 NoP7/B1 21 39 0.57 0.49 0.81 NoP8 21 39 0.38 0.41 0.52 No

Faroqi-Shah and Thompson (Fig. 2)*

[21]English A F

B1/P7 L mca 15 15 0.50 0.60 0.43 NoB2 L mca 15 15 0.53 0.67 0.36 NoB3 L mca 15 15 0.53 0.53 0.64 NoB4 L mca 15 15 0.67 0.53 0.87 No

6Y.Faroqi-Shah

/Brain

Research

Bulletin

74(2007)

1–13Table 1 (Continued )

Reference Language Task/stimuli Patient Lesion # Stimuli Accuracy p One-tail R < I? p < 0.05

R I R I

B5 L mca 15 15 0.40 0.53 0.36 NoB6/P1 L mca 15 15 0.33 0.33 0.65 NoB7 L mca 15 15 0.47 0.33 0.87 NoB8/P3 L mca 15 15 0.47 0.93 0.01 YesB9 L mca 15 15 0.40 0.53 0.43 NoB10 L mca, basal ganglia 15 15 0.87 0.80 0.83 No

Penke and Westerman, p. 568 (Fig. 2)[43]

Dutch B MG L 67 51 1.00 0.60 1.00 NoCO L frontal 30 30 1.00 0.72 1.00 NoHU L dorsal mca 30 30 1.00 0.82 1.00 NoZ 67 51 0.92 0.58 1.00 NoKE L internal capsule 30 30 0.97 0.74 1.00 NoAR L fronto temporal 30 30 0.93 0.74 0.99 NoJW L 67 51 0.98 0.82 1.00 NoKL L mca 30 30 0.97 0.90 0.94 NoNI L mca 30 30 0.97 0.93 0.88 NoAD L mca 67 51 0.92 0.91 0.67 NoAN L 67 51 0.93 0.96 0.35 NoRE 30 30 0.81 0.81 0.63 No

Penke et al., p. 228 (Fig. 1) [45] German B

MT L frontal area of mca 39 39 1.00 0.35 1.00 NoPB L mca 39 39 1.00 0.50 1.00 NoMB L frontal area of mca 39 39 1.00 0.65 1.00 NoKH L mca 39 39 1.00 0.91 1.00 NoWW L mca 39 39 0.99 0.90 1.00 NoHR L mca 39 39 0.99 0.71 1.00 NoJZ L mca 39 39 0.99 0.57 1.00 NoAH L frontal area of mca 39 39 0.91 0.57 1.00 NoGB L mca incl. Broca’s 39 39 0.85 0.80 0.81 NoMJ L 39 39 0.89 0.98 0.18 NoFW L 39 39 0.60 0.65 0.41 No

Tsapkini et al., p. 276 [55] Greek C SK L temporal, internal capsule 11 8 0.27 0.625 0.14 NoLaiacona and Caramazza, p. 116

(Fig. 5), p. 117 (Fig. 6) [35]Italian B MR Frontoparietal, ant. and sup. temporal 204 156 0.67 0.52 1.00 Yes

Balaguer, p. 217–218, p. 220 (Fig. 4)[6]$

Spanish BJM L fronto temporal subcortical, Rt.

periventricular34 30 0.824 0.437 1.00 No

MP L mca 28 25 0.90 0.524 1.00 No

Catalan BJM 34 39 0.912 0.713 1.00 NoMP 35 18 0.80 0.417 1.00 No

Mean 0.59 0.51

Standard deviation 0.31 0.26

Standard error of mean 0.04 0.03

Abbreviations: A, fill blank sentence completion; ant., anterior; B, multiple choice sentence completion; C, response to question; F, frequency matching of regular–irregular; I = irregular; L, left hemisphere; MCA,middle cerebral artery; P, phonological matching of regular–irregulars (syllable structure); post., posterior; R, regular; R < I, accuracy of regulars is less than irregulars; Rt., right; sup., superior; $, past and presentelicited; (*) three participants overlapped in these two studies and their corresponding codes are listed.Disclaimer: In some cases, p-values may vary marginally from those reported by authors since raw data were computed from figures.

Y.Faroqi-Shah/B

rainR

esearchB

ulletin74

(2007)1–13

7Table 2Studies of word repetition sorted by language and chronological order

Author/data location Language stimuli Patient Lesion # Stimuli Accuracy % p One-tail R<I? p < 0.05

R I R I

Badeckar and Caramazza, p. 295(Table 9) [3]

English F FM L post. inf. frontal, inf.parietal, ant. temporal

50 50 0.56 0.74 0.00 Yes

Kohn and Melvold, p. 337 [33] EnglishBD L temporo-parietal, R lacunar 10 10 0.60 0.90 0.15 NoJW L frontal–parietal temporal 10 10 0.50 0.90 0.07 No

Tsapkini et al., p. 278 [55] Greek SK L temporal, internal capsule 43 43 1.00 0.98 1.00 No

Bird et al., p. 507 (Fig. 1), p. 509(Fig. 2) [8]

English F BB L 44 44 0.13 0.32 0.03 YesIJ L 44 44 0.15 0.35 0.04 YesIB L frontal temporal parietal 44 44 0.56 0.82 0.00 YesDE L mca 44 44 0.15 0.65 0.00 YesAB L mca 44 44 0.58 0.82 0.01 YesMB L frontoparietal, R frontal 44 44 0.08 0.20 0.00 YesRT L 44 44 0.75 0.95 0.30 NoVC L frontoparietal 44 44 0.74 0.82 0.01 YesGN L frontal including Broca’s,

occipital44 44 0.76 0.94 0.00 Yes

JL L frontal temporal parietal 44 44 0.66 0.96 0.11 No

English P F

BB 34 34 0.25 0.30 0.47 NoIJ 34 34 0.25 0.18 0.87 NoIB 34 34 0.90 0.89 0.79 NoDE 34 34 0.59 0.50 0.83 NoAB 34 34 0.70 0.80 0.29 NoMB 34 34 0.20 0.25 0.41 NoRT 34 34 0.91 0.91 0.66 NoVC 34 34 0.91 0.85 0.87 NoGN 34 34 0.81 0.85 0.50 NoJL 34 34 0.98 0.92 0.94 No

Lambon Ralph et al., p. 112 (Table 6)[36]

English F AB L 34 34 0.85 0.97 0.10 NoGN L mca, parietal 34 34 0.82 0.94 0.13 NoJL 34 34 0.91 0.91 0.66 NoGD L 34 34 0.97 0.97 0.75 NoPG L subarachanoid hem. 34 34 0.97 0.94 0.88 NoDC 34 34 1.00 0.97 1.00 NoJS 34 34 0.94 0.88 0.90 NoDM 34 34 1.00 0.91 1.00 No

Mean 0.66 0.76

Standard deviation 0.30 0.26

Standard error of mean 0.05 0.04

Abbreviations: ant., anterior; F, frequency matching of regular–irregular; I = irregular; L, left hemisphere; mca, middle cerebral artery; P, phonological matching of regular–irregulars (syllable structure); post.,posterior; R, regular; R < I, accuracy of regulars is less than irregulars; Rt., right; sup., superior.Disclaimer: In some cases, p-values may vary marginally from those reported by authors since raw data were computed from figures.

8 Y. Faroqi-Shah / Brain Research Bulletin 74 (2007) 1–13

Fig. 2. Meta-analysis of sentence completion data showing proportion ofpatients with: significantly lower accuracy of regulars compared to irregulars(bt

Epwmamsftn(wEuf

3

sctislbw

Fscr

ttc

Eafuv

3

a3bdmrltTc

TR

FTFFONN

R < I), irregulars compared to regulars (R > I), and no significant differencesetween accuracies of regulars and irregulars (R–I ns) (Fisher-exact test, one-ailed p < 0.05) (N = 78).

nglish speakers. The number of English speakers who showedoorer accuracy for irregulars is also 9 (out of a total of 26ho showed this pattern). Seven (out of a total of 13) reportsatched regulars and irregulars for frequency of occurrence,

t least for a subset of the stimuli, and only two studies hadatched phonological aspects of regular and irregular verbs,

uch as syllable structure. The number of stimuli used rangedrom 8 to 204 [35,55]. To summarize, over half of the sen-ence production datasets are from English speakers. An equalumber of English speaking patients showed opposite patternsregulars < irregulars and irregular < regulars). If all patients whoere worse on regulars were considered, a majority (9/10) werenglish speakers. Further, only half of the studies matched stim-li for frequency and an even smaller number matched stimulior phonological complexity.

.2. Repetition

Word repetition included 32 data sets from 26 different apha-ic patients (see Table 2). A majority of these data sets (N = 28)ame from the same group of authors [8,36] whose focus waso demonstrate that variations in stimulus characteristics cannfluence the existence of a dissociation. Nine datasets showed
ignificantly lower accuracy of regular verbs compared to irregu-ar verbs. The remaining 23 datasets had no significant differenceetween regulars and irregulars as shown in Fig. 3. No patientsith significantly better performance on irregulars compared
3

o

able 3egular–irregular verb performance for various lesions

# Patients # Datasets R <

rontal excl. temporal 9 12 5emporal excl. frontal 2 3rontal + parietal + temporal 5 12 3rontal + temporal 3 3 1ther (subcortical, subarachnoid, etc.) 12on-specific description (L mca) 35o lesion information 9

ig. 3. Meta-analysis of repetition data showing proportion of patients with:ignificantly lower accuracy of regulars compared to irregulars (R < I), irregularsompared to regulars (R > I), and no significant differences between accuracies ofegulars and irregulars (R–I ns) (Fisher-exact test, one-tailed p < 0.05) (N = 22).

o regulars were reported. There was a high positive correla-ion between the accuracies of regulars and irregulars, whenomputed across all datasets (rs = 0.8, p < 000).

Other variables. Word repetition data are primarily fromnglish speakers, and only one of the 32 datasets came fromnon-English speaker [55]. Four out of six reports had some

requency matching between regulars and irregular verb stim-li, while only one study reported matching of phonologicalariables.

.3. Lesion information

The results of the lesion analysis are reported in Table 3nd Fig. 4. As is evident from Table 3, lesion data from only1 patients could be entered into the analysis of dissociationsecause lesion information was either not reported (N = 9) orescribed in non-specific terms, such as left hemisphere or leftiddle cerebral artery (N = 35). Of particular interest to the neu-

oanatomical hypotheses being evaluated are patients whoseesions were exclusively frontal (or frontoparietal) without anyemporal lobe involvement and vice versa (N = 15 datasets).able 3 reveals that no specific dissociation patterns were asso-iated with either of these lesion sites.

.4. Inter-test consistency

As mentioned earlier, 19 individuals contributed more thanne dataset. A comparison of regular–irregular performance

I R > I R = I, ns Patients

6 1 FCL, MB, VC, RC, PG, CO, MT, AH, GN3 SK, BD9 FM, JW, IB, JL, MR

2 RBA, AR, JM

P1-P7, RE, Z

Y. Faroqi-Shah / Brain Research

aitfppa([w(Tawtsv

4

vws(lrwsits

4v

vuwss

a(TrrtgFtawrTeoa

twB[hvwtofiwccpaiobrdouica

iirgasatir

Fig. 4. Proportion of patients with various lesion sites.

cross the different datasets of each of these patients revealedntra-individual variations in dissociation patterns. In 13 cases,he profile changed from regular < irregular to no significant dif-erence between regulars and irregulars. A majority of theseatients were reported by the same group of authors whoserimary focus was to demonstrate how dissociation patternsre vulnerable to stimulus manipulations [8,36]. A patientB8/P3) who participated in studies by two different authors21,22], was significantly worse on regulars in one studyhile being significantly worse on irregulars in another study

Table 1). These two studies used different task variations.he performance of five other patients remained unchangedcross multiple datasets. To summarize, these 19 individualsho contributed multiple datasets, demonstrate poor inter-

est consistency. This indicates that dissociation patterns areusceptible change across testing sessions and experimentalariations.

. Discussion

This meta-analysis of published research on regular–irregularerb production in individuals with non-fluent/Broca’s aphasiaas conducted to address three questions. (1) Is there a con-

istent pattern of dissociation between regulars and irregulars?2) Are specific dissociation patterns associated with specificesions? In particular, is the pattern of worse performance onegulars associated with frontal lobe lesions? More specifically,e compared the predictions of the declarative-procedural [59],

emantic-phonology [57] and two-stage [5] models, and (3) ist possible to delineate any variables that influence dissocia-ion patterns? These questions are addressed in the followingections.

.1. Is there a dissociation between regular and irregularerbs?

The most evident aspect of the data presented is the greatariability across patients in the production of regular and irreg-
lar verbs, although all patients were described as individualsith non-fluent/Broca’s aphasia. Combined across the 110 data
ets of sentence production and word repetition, all three pos-ible patterns were observed: no difference between regulars

omd[

Bulletin 74 (2007) 1–13 9

nd irregulars (65 datasets), worse performance on irregulars26 datasets) and worse performance on regulars (9 datasets).he most frequent pattern in both sentence production and word

epetition tasks was that of no significant difference betweenegulars and irregulars (Figs. 2 and 3). It is also noteworthyhat only five datasets out of a total of 110 had a difference ofreater than 50% between regulars and irregulars (sentences:CL, GN, MT, PB; repetition: DE). Among these five datasets,

here was no consistent pattern. In other words, classical dissoci-tions, the strongest form of neuropsychological evidence [52],ere hardly found. Moreover, a significantly high positive cor-

elation was found between regular and irregular verb accuracy.he observed heterogeneity, predominant pattern of no differ-nce, and high positive correlation, argue against the presencef a clear and consistent pattern of dissociation between regularsnd irregulars in Broca’s aphasia.

It is evident that the declarative/procedural [59] and seman-ic/phonology [31,32,42,56,57] models, both of which predictorse performance on regular compared to irregular verbs inroca’s aphasia, are not supported. The semantics/phonology

31,32] model, assumes that individuals with Broca’s aphasiaave phonological deficits and hence difficulty with regularerb production. Since the presence of phonological deficitsas not documented by most authors, it may be preliminary

o reject its claims on the basis of lack of evidence. In a studyf two fluent and two non-fluent aphasic individuals with con-rmed phonological deficits, repetition of regular verbs wasorse than irregulars in all four patients [33]. In another study,

ontrolling for phonological difficulty eliminated prior disso-iations between regulars and irregulars in non-fluent aphasicatients [8]. Hence, it appears that phonological complexity isstimulus artifact that needs to be controlled in future stud-

es of regular–irregular verb processing. Given that a majorityf data sets found no regular–irregular differences, the data areest accommodated by viewpoints that assume more similaritiesather than differences between regular and irregular verb pro-uction, such as the two-stage model [5,6,37]. The second stagef the two-stage model is assumed to be common to both irreg-lar and regular verbs by involving retrieval of morphosyntacticnformation, such as tense [5,6]. Further, this second stage islaimed to be processed by the LIFG, a brain area whose lesionsre commonly associated with Broca’s aphasia.

The neuroanatomical models that were developed on thedea that regulars are impaired in Broca’s aphasia have largelygnored the fact that some aspects of regular morphology areelatively spared in Broca’s aphasia [41,57]. For example, pro-ressive forms of verbs (kicking, sleeping), verbs that inflect togree with the preceding noun (The cat sleeps versus The catsleep), and plural noun morphology (dogs) are all regular andre produced with relatively high accuracy [23,24]. This meanshat models that make a binary distinction between regulars andrregulars while failing to incorporate the whole spectrum ofegular morphological operations are inadequate. Any model
f the neurocognitive differences between regular and irregularorphology is incomplete if it does not account for the relative
ifferences among these various regular morphological forms35,53].

1 search

4d

ftLhiftwrwwlsulcdptsra

st[ttiodaffnachB

lytbtta

4v

q

tduiotrfa

cTadcsbicvoipnfbcmIitfqtsstca

idpm

5

dciation between regular and irregular verbs in persons describedas non-fluent or Broca’s aphasia, at least for sentence and wordrepetition tasks reviewed in this meta-analysis3; (2) the most fre-

0 Y. Faroqi-Shah / Brain Re

.2. Are specific lesions sites associated with specificissociation patterns?

Not surprisingly, lesions of the left hemisphere were reportedor all the participants. A variety of lesions, including cor-ical and subcortical regions, were reported. Given that theIFG is frequently credited with the mental grammar andence production of regular verbs [41,59], the most interest-ng group constitutes patients with exclusively frontal lobe (orrontal–parietal) lesions. We are making the liberal assump-ion that the lesion was inclusive of the LIFG in all patientsith reported frontal lobe lesions. Considerable variability in

egular–irregular performance is observed across the 17 patientsho had some form of frontal lobe lesion (see Table 3),ith approximately equal numbers of datasets showing regu-

ars worse and better than irregulars. Hence, the data do notupport a special connection between LIFG lesions and a reg-lar verb deficit. The three patients who had temporal lobeesions with the sparing of frontal lobe showed no signifi-ant differences between regulars and irregulars. Recall that theeclarative/procedural and semantic/phonology models predictoor performance on irregulars in temporal lobe lesions, whilehe two-stage model predicts poor regular verb performance foruch patients. To summarize, there seems to be no evident cor-elation between any pattern of performance and specific frontalnd temporal lesion sites.

These complex lesion-deficit findings are actually con-istent with neuroimaging studies of regular–irregular verbshat have found a wide variety of activation patterns16,17,28,30,49,54,56]. No two-brain imaging studies agree onhe regions activated for regular and irregular verbs, perhaps dueo methodological variations. In other words, neuropsycholog-cal and neuroimaging data suggest that the neural generationf regulars and irregulars is far more complex than that pre-icted by any current neuroanatomical model. One similaritycross most brain imaging studies is the reported activation ofrontal regions for both regular and irregular verbs. As for leftrontal lobe involvement, the two-stage model [5] is the onlyeuroanatomical model that specifically implicates both regularnd irregular verb processing for this region. Hence, one mayonclude (though not without reservations) that the finding ofigh correlation between regular and irregular verb scores inroca’s aphasia fits best with the two-stage model.

A cautionary note about the present lesion analysis is thatess than half of all patients could contribute to the lesion anal-sis because lesion data in other studies were not reported inerms that were specific enough for inclusion. It is quite possi-le that the null result may be because the lesion descriptions areoo vague and non-specific to reveal any correlations. Perhapshis question needs to be addressed using a more fine-grainedpproach, such as voxel-based lesion analysis [56].

.3. Can dissociations be manipulated by experimental
ariables?
At least three distinct points can be made regarding thisuestion. First, a non-trivial finding of this meta-analysis was

q

e

Bulletin 74 (2007) 1–13

hat a majority of patients who contributed more than oneataset showed variations in the relative performance of reg-lars and irregulars across the datasets (Section 3.4). Thisntra-individual variability questions the robustness of previ-usly reported “regular–irregular dissociations” and highlightshe importance of demonstrating test–retest consistency in neu-opsychological data, especially when such data are used torame theories about the normal neural processing of regularnd irregular verbs [56,59].

Secondly, the data suggest that at least some reported disso-iations may have been an artifact of stimulus manipulations.his point is especially made by a few studies in whichpparent dissociations between regular and irregular verb pro-uction were eliminated when verbs with matched phonologicalomplexity (the number of syllables, the presence of con-onant clusters) were used [8,36]. It was emphasized that,ecause a majority of regularly inflected verbs are phonolog-cally complex in English, failure to control for phonologicalomplexity disproportionately affects the production of regularerbs. Phonological factors seem to influence the occurrencef regular–irregular dissociations in another manner. Aphasicndividuals with phonological deficits have a greater difficultyroducing regular verb inflections in English, even when they doot have Broca’s aphasia [33]. Hence, the pattern of worse per-ormance on regulars is not invariably linked to Broca’s aphasia,ut may be influenced by phonological deficits [57] and stimulusharacteristics [8,36]. There may also be an interaction betweenodality (reading versus repetition) for phonological variables.

n a previous study, matching phonological complexity in a read-ng task (links versus lynx) still showed a worse performance forhe morphologically complex word (links) [3,4]. A variety ofrequency measures including affix frequency [11] lexeme fre-uency [20], and stem frequency [45] have also been showno influence accuracy of verb morphology. As Tables 1 and 2how, a majority of the studies did not control frequency mea-ures between regular and irregular stimuli, thereby weakeninghe findings. Task variations are another factor that need carefulonsideration as is evident with patient B6/P1 [21,22] in Table 1nd also in other studies [8,34].

Finally, the reviewed studies had at least two methodolog-cal weaknesses that are likely to inflate the occurrence ofissociations: use of experimental tasks that produced ceilingerformance in unimpaired controls, and the failure to use aatched control group [13,14,38,52].

. Conclusions and future research considerations

The following statements can be made on the basis of the 110atasets reviewed: (1) there was no consistent pattern of disso-

uent pattern was that of no difference in performance; (3) the

3 The semantics/phonology model is primarily based on auditory same differ-nt discrimination and lexical priming [56].

search

mtawpvm

cmdsaisApaAirsicatot[

epmrtasnbmculfiptiof“

ru

d

nam

tiafArnopd5Bit[mfi

psrBbplagtImtainiflo

C

A

J


agnitude of difference between regular and irregular produc-ion accuracy was minimal, with only five individuals showingdifference of greater than 50%; (4) a wide variety of lesionsere reported; (5) lesion sites were not associated with specificerformance patterns in any consistent manner; and (6) inter-testariability was found in nearly two-thirds of patients who hadultiple datasets.The absence of a consistent dissociation and lesion-deficit

orrelation can be interpreted in different ways. First, the resultsay be taken at face value to mean that there is indeed no

issociation between regulars and irregulars in Broca’s apha-ia. Regular and irregular verbs may fail to dissociate becausecomplex network of multiple, overlapping brain regions are

nvolved in producing both verb types, and this was demon-trated by brain imaging research [16,17,28,30,32,49,54,56].lternatively, these verbs may fail to dissociate because thesycholinguistic dichotomy between regulars and irregularsssumed by aphasiologists is overly simplistic [26,54,22,19].rguments for a gradation between regulars and irregulars

nclude the presence of subregularities among irregulars (sang,ang, drank versus slept, kept, wept) [9], affix-like patterns inome irregulars (kept → kep + t) [22,26], presence of affixesn irregular verbs of other languages [6,43], and “morphologi-ally simple” behavior of high frequency regulars [1]. As someuthors have previously pointed out, regular and irregular pastense verbs have more commonalities (morphosemantic notionf past tense, syntactic structure of the sentence, experimen-al task demands, etc.) than the one morphological difference5,6,37].

As for the three neuroanatomical viewpoints presentedarlier, the existing data are incompatible with the declarative-rocedural model [41], but fairly consistent with the two-stageodel [5]. The two-stage model assumes some degree of neu-

oanatomical overlap between regulars and irregulars and hencehe failure of regulars and irregulars to completely dissoci-te. Some aspects of the results may be consistent with theemantic-phonology model, especially if one makes the con-ection between regulars and phonological skills rather thanetween regulars and frontal lobe. Another neuroanatomicalodel that is a viable alternative is the memory, unification,

ontrol (MUC) framework [25], which allocates the role ofnification to the LIFG. Unification is defined as combininginguistic elements into larger units and it is recognized that uni-cation operations occur in parallel in semantic, syntactic, andhonological domains. Although the author does not explicitlyalk about morphology, it can be seen how both regulars andrregulars involve the same semantic and syntactic unificationperations, differing only in morphological unification. There-ore LIFG lesions would impair both regular and irregular verbunification” operations.4

There are several important considerations for futureesearch. First, the analysis underlines the relevance of metic-lous control and documentation of confounding variables,

4 In its present form, the MUC framework [61] is drawn from comprehensionata.

av

R

Bulletin 74 (2007) 1–13 11

umber of stimuli, task demands, inter-session consistency, andssociated deficits when using neuropsychological data to for-ulate theories of the neural organization of language.Secondly, a frequent logical fallacy when drawing interpre-

ations about normal language function from studies of aphasicndividuals is the over-simplified assumption that all Broca’sphasic individuals have a lesion of the Broca’s area, and there-ore observed deficits reflect functions of a normal Broca’s area.

fine-grained documentation of the extent of lesion is war-anted in every study that makes specific claims about normaleural architecture. There are previous instances when a studyf Broca’s aphasic individuals is cited elsewhere as a study ofatients with lesions of LIFG, even though the original authorsid not report LIFG lesions (e.g. [5, p. 883]). In fact, only0–60% of patients with lesions to Broca’s area have persistentroca’s aphasia and about 15% of right-handed Broca’s aphasic

ndividuals do not have lesions of Broca’s area [18]. Also givenhe heterogeneity of the definition of Broca’s area across studies39], neuroanatomical models that are based on patient perfor-ance need to explicitly define their “Broca’s area” and providene-grained lesion information.

Third, future models also need to incorporate a more com-lete range of morphological operations, such as noun plurals,ubject verb agreement, and non-finite verbs. This is war-anted because recent research suggests that the trouble inroca’s aphasia may not be morphological regularity per se,ut rather an effect of grammatical class [35,53], or mor-hosemantic aspects, such as tense marking [21,23,61]. Finally,anguage specific effects need to be considered since this meta-nalysis revealed a high proportion of English speakers in theroup that performed worse on regular verbs. Currently, onlyhe two-stage model accommodates language specific patterns.t is noteworthy that recent neuroanatomical hypotheses areoving in the right direction and becoming increasingly sophis-

icated in their empirically driven descriptions of processesnd neural regions involved in the generation of regular andrregular verbs [5,33,57]. To conclude, this meta-analysis doesot claim that regular and irregular verbs cannot dissociaten non-fluent/Broca’s aphasia. Rather, the finding is that non-uent/Broca’s aphasia is not associated with a consistent patternr direction of dissociation.

onflict of interest

None.

cknowledgements

The author wishes to thank Rachel Mont, Adrianna Naim andennifer Maultasch for help with data coding. I also thank twononymous reviewers for their helpful comments on an earlierersion of this article.

eferences

[1] M. Alegre, P. Gordon, Frequency effects and the representational status ofregular inflections, J. Mem. Lang. 40 (1999) 41–61.

1 search

[

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[[

[

[

[

[

[

[[

[

[

[

[

[

[

2 Y. Faroqi-Shah / Brain Re

[2] H. Baayen, R. Schreuder, War and peace: morphemes and full forms in anoninteractive activation parallel dual-route model, Brain Lang. 68 (1999)27–32.

[3] W. Badeckar, A. Caramazza, The analysis of morphological errors in a caseof acquired dyslexia, Brain Lang. 32 (1987) 278–305.

[4] W. Badeckar, A. Caramazza, Morphological composition in the lexicaloutput system, Cogn. Neurosci. 8 (1991) 335–367.

[5] R.D. Balaguer, A. Rodriguez-Fornells, M. Rotte, J. Bahlmann, H.J. Heinze,T.F. Munte, Neural circuits subserving the retrieval of stems and grammat-ical features in regular and irregular verbs, Hum. Brain Map. 27 (2006)874–888.

[6] R.D. Balaguer, Regular and irregular morphology and is relationship withagrammatism: evidence from two Spanish-Catalan bilinguals, Brain Lang.91 (2004) 212–222.

[7] E. Bates, M. Appelbaum, J. Salcedo, A.P. Saygin, L. Pizzamiglio,Quantifying dissociations in neuropsychological research, J. Clin. Exp.Neuropsychol. 25 (2003) 1128–1153.

[8] H. Bird, M.A. Lambon Ralph, M.S. Seidenberg, J.L. McClelland, K.Patterson, Deficits in phonology and past-tense morphology: what’s theconnection? J. Mem. Lang. 48 (2003) 502–526.

[9] J. Bybee, D.I. Slobin, Rules and schemas in the development and use ofthe English past tense, Language 58 (1982) 265–289.

10] J. Bybee, Regular morphology and the lexicon, Lang. Cogn. Processes 10(1995) 425–455.

11] J.G. Centeno, L.K. Obler, Agrammatic verb errors in Spanish speakers andtheir normal discourse correlates, J. Neuroling. 14 (2001) 349.

12] H. Clahsen, M. Almazan, Syntax and morphology in William’s syndrome,Cognition 68 (1998) 167–198.

13] J.R. Crawford, P.H. Garthwaite, Detecting dissociations in single-case stud-ies: type I errors, statistical power and the classical versus strong distinction,Neuropsychologia 44 (2006) 2249.

14] J.R. Crawford, P.H. Garthwaite, C. Gray, Wanted: fully operational defini-tions of dissociations in single case studies, Cortex 39 (2003) 357–370.

15] A.R. Damasio, Aphasia. New Engl. J. Med. 326 (1992) 531–539.16] M.H. Davis, F. Meunier, W.D. Marslen-Wilson, Neural responses to mor-

phological, syntactic, and semantic properties of single words: an FMRIstudy, Brain Lang. 89 (2004) 439.

17] R.P. Dhond, K. Marinkovic, A.M. Dale, T. Witzel, E. Halgren, Spa-tiotemporal maps of past-tense verb inflection, Neuroimage 19 (2003) 91–100.

18] N.F. Dronkers, J.K. Shapiro, B. Redfern, R.T. Knight, The role of Broca’sarea in Broca’s aphasia, J. Clin. Exp. Neuropsychol. 14 (1992) 52–53.

19] J. Druks, Morpho-syntactic and morpho-phonological deficits in the pro-duction of regularly and irregularly inflected verbs, Aphasiology 20 (2006)993–1017.

20] Y. Faroqi-Shah, C.K. Thompson, Semantic, lexical, and phonological influ-ences on the production of verb inflections in agrammatic aphasia, BrainLang. 89 (2004) 484–498.

21] Y. Faroqi-Shah, C.K. Thompson, Verb inflections in agrammatic aphasia:encoding of tense features, J. Mem. Lang. 56 (2007) 129–151.

22] S. Fix, C.K. Thompson, Morphophonological structure and agrammaticregular/irregular past-tense production, Brain Lang. 99 (2006) 166–167.

23] N.a. Friedmann, Y. Grodzinksy, Tense and agreement in agrammatic pro-duction: pruning the syntactic tree, Brain Lang. 56 (1997) 397–425.

24] H. Goodglass, J.A. Christiansen, R. Gallagher, Comparison of morphologyand syntax in free narrative and structured tests: fluent versus nonfluentaphasics, Cortex 29 (1993) 377–407.

25] P. Hagoort, On Broca, brain, and binding: a new framework, Trends Cogn.Sci. 9 (2005) 416–423.

26] M. Halle, A. Marantz, Distributed morphology and the pieces of inflection,in: K. Halle, S.J. Keyser (Eds.), The View from Building 29: Essays inLinguistics in Honor of Sylvian Bromberger, MIT Press, Massachusetts,1993, pp. 111–175.

27] G. Hickok, D. Poeppel, Towards a functional neuroanatomy of speechperception, Trends Cogn. Sci. 4 (2000) 131–138.

28] P. Indefrey, C. Brown, P. Hagoort, H. Herzog, M. Sach, R.J. Seitz, A PETstudy of cerebral activation patterns induced by verb inflection, NeuroIm-age 5 (1997) S548.

[

Bulletin 74 (2007) 1–13

29] A.L. Inglis, Is ‘time’ up for the irregularity of tense in aphasia, in relationto semantics and the temporal lobe? Brain Lang. 95 (2005) 189.

30] J.J. Jaeger, A.H. Lockwood, D.L. Kemmerer, J. Van Valin, B.W. Mur-phy, H.G. Khalak, A positron emission tomographic study of regularand irregular verb morphology in English, Language 72 (1996) 451–493.

31] M.F. Joanisse, M.S. Seidenberg, Impairments in verb morphology afterbrain injury: a connectionist model, Proc. Natl. Acad. Sci. U.S.A. 96 (1999)7592–7597.

32] M. Joanisse, M. Seidenberg, Imaging the past: neural activation in frontaland temporal regions during regular and irregular past tense processing,Cogn. Affect. Behav. Neurosci. 5 (2005) 282–296.

33] S. Kohn, J. Melvold, Effects of morphological complexity on phonologi-cal output deficits in fluent and nonfluent aphasia, Brain Lang. 73 (2000)323–346.

34] P. Kok, A. van Doorn, H. Kolk, Inflection and computational load in agram-matic speech, Brain Lang. 102 (2007) 273–283.

35] M. Laiacona, A. Caramazza, The noun/verb dissociation in languageproduction: varieties of causes, Cogn. Neuropsychol. 21 (2004) 103–123.

36] M.A. Lambon Ralph, N. Braber, J.L. McClelland, K. Patterson, Whatunderlies the neuropsychological pattern of irregular > regular past-tenseverb production? Brain Lang. 93 (2005) 106–119.

37] A. Laudanna, Regular versus irregular inflection: a question of levels,Behav. Brain Sci. (1998) 1028–1029.

38] K.R. Laws, T.M. Gail, V.M. Leeson, J.R. Crawford, When is categoryspecific in Alzheimer’s disease? Cortex 41 (2005) 452–463.

39] R. Lindenberg, H. Fangerau, R.J. Seitz, “Broca’s area” as a collective term?Brain Lang. 102 (2007) 22–29.

40] http://www.matforsk.no/ola/fisher.htm.41] W.D. Marslen-Wilson, L.K. Tyler, Dissociating types of mental computa-

tion, Nature 387 (1997) 592–594.42] W.D. Marslen-Wilson, L.K. Tyler, Morphology, language and the brain:

the decompositional substrate for language comprehension, Philos. Trans.R. Soc. B 362 (2007) 823–836.

43] M. Penke, G. Westermann, Broca’s area and inflectional morphology:evidence from Broca’s aphasia and computer modeling, Cortex (2006)563–576.

44] M. Penke, H. Weyerts, M. Gross, E. Zander, T.F. Munte, H. Clahsen, Howthe brain processes complex words: an event-related potential study ofGerman verb inflections, Cogn. Brain Res. 6 (1997) 37–52.

45] M. Penke, U. Janssen, M. Krause, The representation of inflectional mor-phology: evidence from Broca’s aphasia, Brain Lang. 68 (1999) 225–232.

46] S. Pinker, M.T. Ullman, The past and future of the past tense, Trends Cogn.Sci. 6 (2002) 456–463.

47] S. Pinker, The Language Instinct, William Morrow, New York, 1994.48] D. Rumelhart, J. McClelland, On learning the past tenses of English

verbs: implicit rules or parallel distributed processing, in: D. Rumel-hart, J. McClelland (Eds.), Parallel Distributed Processing: Explorationsin the Microstructure of Cognition, MIT Press, Cambridge, MA,1986.

49] M. Sach, R.J. Seitz, P. Indefrey, Unified inflectional processing of regularand irregular verbs: a PET study, Neuroreport 15 (2004) 533–537.

50] E.M. Saffran, R.S. Berndt, M.F. Schwartz, The quantative analysis ofagrammatic production: procedure and data, Brain Lang. 37 (1989)440–479.

51] S. Seigel, J.N. Castellan, Nonparametric Statistics for the Behavioral Sci-ences, second ed., McGraw Hill, Boston, 1988.

52] T. Shallice, From Neuropsychology to Mental Structure, Cambridge Uni-versity Press, Cambridge, UK, 1988.

53] K. Shapiro, A. Caramazza, Grammatical processing of nouns and verbs inthe left frontal cortex? Neuropsychologia 41 (2003) 1189–1198.

54] L. Stockall, A. Marantz, A single route, full decomposition model of
morphological complexity: MEG evidence, Mental Lexicon 1 (2006) 85–123.
55] K. Tsapkini, G. Jarema, E. Kehayia, A morphological processing deficitin verbs but not in nouns: a case study in a highly inflected language, J.Neuroling. 15 (2002) 265–288.

search

[

[

[

[

[


56] L.K. Tyler, W.D. Marslen-Wilson, E.A. Stamatakis, Differentiating lexicalform, meaning, and structure in the neural language system, Proc. Natl.Acad. Sci. U.S.A. 102 (2005) 8375–8380.

57] L.K. Tyler, E.A. Stamatakis, B. Post, B. Randall, W. Marslen-Wilson, Tem-poral and frontal systems in speech comprehension: an FMRI study of past
tense processing, Neuropsychologia 43 (2005) 1963–1974.
58] M.T. Ullman, R. Bergida, M. O’ Craven, Distinct fMRI activation pat-terns for regular and irregular past tense, in: Paper Presented at the ThirdInternational Conference on Functional Mapping of the Human Brain,Copenhagen, Denmark, 1997.

[

Bulletin 74 (2007) 1–13 13

59] M.T. Ullman, S. Corkin, M. Coppola, G. Hickok, A neural dissociationwithin language: evidence that the mental dictionary is part of declara-tive memory, and that grammatical rules are processed by the proceduralsystem, J. Cogn. Neurosci. 9 (1997) 266–276.

60] M.T. Ullman, R. Pancheva, T. Love, E. Yee, D. Swinney, G. Hickok,
Neural correlates of lexicon and grammar: evidence from the produc-tion, reading, and judgment of inflection in aphasia, Brain Lang. 93 (2005)185.
61] M. Wenzlaff, H. Clahsen, Tense and agreement in German agrammatism,Brain Lang. 89 (2004) 57–68.

Are regular and irregular verbs dissociated in non-ﬂuent ...

Documents