Testing the limits of language production in long-term survivors of major stroke: A psycholinguistic and anatomic study

Abstract

Background: There is still a dearth of information about grammatical aspects of language production in aphasia.

Aims: Making novel use of methods of elicited production aimed at testing the limits of competence, we studied three cases of chronic aphasia, stemming from major stroke. We asked: (1) Whether the elicited production method reveals sparing of language abilities not readily evidenced in spontaneous utterances or on conventional aphasia tests. (2) Which language production abilities survive damage to both Broca's region and Wernicke's region?

Materials & Procedures: Targeted words, morphological and syntactic structures were elicited by sentence completion with supporting linguistic and visual context. Targets were never modelled during the procedure. For verbs, visual and auditory contexts emphasise completed actions, targeting past tense forms. Lesion description was based on structural MRI scans.

Outcomes & Results: The three participants showed partially spared ability to produce nouns, adjectives, and verb stems in context. The elicitation method proved more productive in some cases than picture prompts or sentence prompts. Past tense inflections were usually omitted. Hence stems and inflections were dissociable. Two participants showed partial success with the passive, and no participant produced a full relative clause, including the relative pronoun, but two produced reduced forms of subject relatives. Partial sparing of production capability in these cases points to the likely importance of portions of the left hemisphere remote from Broca and Wernicke regions.

Conclusions: This application of elicited production methodology demonstrates possibilities of lexical, morphological, and syntactic production not evident in spontaneous utterances or by conventional aphasia tests. Some lexical and grammatical capabilities survived massive damage to both anterior and posterior portions of the left hemisphere.

Keywords:

• Aphasia
• Grammatical production
• Recovery of production

We thank the participants for their completion of a lengthy project with patience and good humour. We also thank two anonymous reviewers for their substantial help in improving the clarity of both the psycholinguistic and anatomic aspects. This research was supported in part by grants in support of faculty research from the University of Connecticut Research Foundation. We thank George Wolford for assistance in arranging for an MRI scan at the Dartmouth College Imaging Center. Preparation of the manuscript was partially supported by a grant from the National Institutes of Health, HD-40353 to Haskins Laboratories.

Notes

¹HW was tested with the same materials 5 years prior to the reported testing session. At that previous testing session he produced one of the derived adjective targets (wrinkled), and made another related substitution (“bent” for “twisted”). At this testing session he was also successful at producing “wrinkled” as a past-tense verb.

²A summary of some of the psycholinguistic data (that relating to relative clause elicitation) for one of the participants, HW, was included in an earlier review article which interpreted sentence-processing deficits in aphasia in the light of findings from eye-tracking studies of sentence processing by non-impaired adults, and sentence processing and experimental studies of language in developing children (Crain et al., 2001).

³Homophony across word classes added a control to the design of ELEX. In all instances past-tense verb targets and noun/adjective targets were homophonic pairs, thereby allowing us to rule out one possible source of differences in retrievability that would confound comparisons across word classes (see also Shapiro & Caramazza, 2003). Kemmerer and Tranel (2000) and Bastiaanse et al. (2003) report that homophony between verbs and nouns has a positive influence on verb retrieval.

⁴It has often been pointed out that the interhemispheric distribution of language functions varies widely from person to person for reasons that are still only partially understood. One known source of variability is handedness. Because BN was judged to have been ambidextrous before her stroke, there is reason to ask whether in recovery she may have benefited from a pre-existing bilateral organisation of language, a condition that sodium amytal studies have shown can be a reality in some (mainly) non-right-handed individuals (Rasmussen & Milner, 1977). BN did indeed show an unexpected amount of recovery of both production and comprehension.