Recovery in deep dysphasia: A model-based approach

Walter Huber *, Irene Ablinger, Stefanie Abel

Section Neurolinguistics, Department of Neurology, University Hospital, RWTH Aachen University, Pauwelsstrasse 30, D-52057 Aachen, Germany

Introduction

The syndrome of deep dysphasia is characterized by an inability torepeat pseudowords and the production of semantic errors in word rep-etition. Several single case studies revealed that phonological decodingmight be outstandingly impaired (cf. Jefferies, Sage, & Lambon Ralph,2007). Recovery of deep dysphasia has only been illustrated in detailfor patient NC (Martin & Saffran, 1992). Dell, Schwartz, Martin, Saf-fran, and Gangnon (1997) tried to simulate NC’s repetition perfor-mance in their connectionist lexical activation model, but it couldnot fit his error pattern as it assumes perfect recognition of auditoryinput.

In a new case study on recovery of deep dysphasia, we intended to col-lect further evidence for the assumption that impaired input processing isthe crucial cause of the impairment. Moreover, we aimed to explainimpairment and psycholinguistic parameter effects in the connectionistsemantic-phonological model (Foygel & Dell, 2000) by adding a phoneticinput level.

Methods

JR is a native speaker of German who worked as a software man-ager. At the age of 46, he suffered from an ischemic posterior MCAinfarct. He was extensively examined at 1, 3, 8 and 10 months postonset.

Errors in naming and repetition were classified according to the taxon-omy of Dell et al. (1997). JR’s error patterns were simulated in the seman-tic-phonological model to determine the naming disorder and to predictword repetition.

In addition, we established an error modality analysis, to disentangleinput and output impairments in repetition. Thus, the source of each errorcould be subclassified as belonging to either expressive or receptive com-ponents of repetition.

Results

At first testing, JR presented with fluent paraphasic and paragrammat-ic speech and a sharp performance contrast between severely impaired rep-etition and fairly preserved reading aloud, while naming on confrontationwas moderately affected. During recovery, performance in naming andword repetition highly improved, while repetition of pseudowordsremained impossible. The evolvement of real word repetition was charac-terized by psycholinguistic parameter effects at different points in time:concreteness before length before frequency (see Table 1).

The connectionist model overpredicted correct responses in word rep-etition as was already demonstrated for NC.

There were only few expressive repetition errors; regarding receptiveerrors, nonwords and null responses decreased significantly while formalerrors became the dominant error type in the course of recovery.

Discussion

JR’s dissociations in performance, the evolvement of psycholinguisticparameter effects, the insufficient computer simulations derived from nam-ing, results from error modality analysis as well as changes of error patternare ample evidence for the primary decoding disorder. We argue that deepdysphasia can be explained by an impairment of phonetic-phonologicalconnections in the extended version of the one-route model of repetitionwith 4 rather than 3 levels of auditory word processing. JR’s improved realword repetition despite persisting failure on pseudowords is accounted forby an increase of both phonetic-phonological and lexical-phonologicalconnection weights.

In conclusion, we present a new case of recovering deep dysphasiawhich cannot be directly explained by the current versions of the con-nectionist lexical activation model (cf. the recent discussion in Schwartz,Dell, Martin, Gahl, & Sobel (2006), and in Dell, Martin, & Schwartz(2007)).

doi:10.1016/j.bandl.2007.07.097

* Corresponding author. Fax: +49 241 8082598.E-mail address: whuber@ukaachen.de (W. Huber).

www.elsevier.com/locate/b&l

Brain and Language 103 (2007) 8–249

References

Dell, G. S., Schwartz, M. F., Martin, N., Saffran, E. M., & Gangnon, D.A. (1997). Lexical access in aphasic and nonaphasic speakers.Psychological Review, 104, 801–838.

Dell, G. S., Martin, N., & Schwartz, E. M. (2007). A case-series test of theinteractive two-step model of lexical access: Predicting word repetitionfrom picture naming. Journal of Memory and Language, 56, 490–520.

Foygel, D., & Dell, G. S. (2000). Models of impaired lexical access inspeech production. Journal of Memory and Language, 43, 182–216.

Jefferies, E., Sage, K., & Lambon Ralph, M. A. (2007). Do deep dyslexia,dysphasia and dysgraphia share a common phonological impairment?Neuropsychologia, 45, 1553–1570.

Martin, N., & Saffran, E. M. (1992). A computational account of deepdysphasia: Evidence from a single case study. Brain and Language, 43,240–474.

Schwartz, M. F., Dell, G. S., Martin, N., Gahl, S., & Sobel, P. (2006). Acase-series test of the interactive two-step model of lexical access:Evidence from picture naming. Journal of Memory and Language, 54,228–264.

Table 1JR’s Repetition of words: influence of meaning, length and frequency

T1 T2 T3 T4Item number correct Item number correct Item number correct Item number correct

Words 55/160 83/160 105/160 129/160Concrete 42/80 51/80 65/80 72/80Abstract 13/80 32/80 40/80 57/80

Concreteness (p-value)* <.001 .002 <.001 .002

1 syll. concrete 23/40 27/40 33/40 36/402 syll. concrete 19/40 24/40 32/40 36/40

Length (p-value)* .186 .321 .500 .644

1 syll. abstract 11/40 20/40 21/40 30/402 syll. abstract 2/40 12/40 19/40 27/40

Length (p-value)* .007 .055 .412 .311

High freq. concrete 22/40 33/40 37/40 37/40Low freq. concrete 20/40 18/40 28/40 35/40

Frequency (p-value)* .328 <.001 .002 .356

High freq. abstract 6/40 18/40 22/40 34/40Low freq. abstract 7/40 14/40 18/40 23/40

Frequency (p-value)* .50 .247 .251 .006

* Fisher’s exact test; freq. = frequency; syll. = syllable.

Abstract / Brain and Language 103 (2007) 8–249 167