Amnesic H.M.'s performance on the language competence test: Parallel deficits in memory and sentence production

Abstract

To test conflicting hypotheses regarding amnesic H.M.'s language abilities, this study examined H.M.'s sentence production on the Language Competence Test (Wiig & Secord, 1988). The task for H.M. and 8 education-, age-, and IQ-matched controls was to describe pictures using a single grammatical sentence containing prespecified target words. The results indicated selective deficits in H.M.'s picture descriptions: H.M. produced fewer single grammatical sentences, included fewer target words, and described the pictures less completely and accurately than did the controls. However, H.M.'s deficits diminished with repeated processing of unfamiliar stimuli and disappeared for familiar stimuli—effects that help explain why other researchers have concluded that H.M.'s language production is intact. Besides resolving the conflicting hypotheses, present results replicated other well-controlled sentence production results and indicated that H.M.'s language and memory exhibit parallel deficits and sparing. Present results comport in detail with binding theory but pose problems for current systems theories of H.M.'s condition.

The authors gratefully acknowledge support from the Samuel A. MacKay Memorial Research Fund and NIH Grant R01 AG 09755. We thank Dr. Suzanne Corkin for permission to study H.M., Kataryna Boese, Alicia Chang, Meghan Gould, Jenifer Taylor, and Diane Marian for general research assistance, Kethera Folger for running the preliminary rating study, and Pamela Crombie and Irwin Stein for providing control participants from the UCLA Cognition and Aging Laboratory and Dr. D. Burke for providing control participants from the Claremont Project on Memory and Aging. Finally, we thank an anonymous reviewer for suggesting “processing difficulty” as an alternate account of the present data (see the Discussion section).

Notes

¹Recent magnetic resonance imaging (MRI) data (Corkin, Amaral, González, Johnson, & Hyman, 1997) indicate a large cerebellar lesion due to H.M.'s use of dilantin for treating epilepsy since 1953, together with possible damage to lateral temporal neocortex that was not due to the original surgery. This possible but at most minimal damage may reflect either an age-linked effect or occurrence of transneuronal degeneration subsequent to H.M.'s 1953 surgical lesion.

²All instances of the label “unintelligible” (N = 25) occurred for H.M., and two factors made H.M.'s speech difficult to decipher: incoherent and difficult to understand syntax (discussed shortly) and poor articulatory quality (undoubtedly related to H.M.'s cerebellar lesion noted earlier).

³By standard convention, the magnitude of infinitely large deficits (as can occur when the controls perform a task with SD = 0) is indicated conservatively as “in excess of 6 SDs.”

⁴In fact, significantly superior TTR performance for H.M. relative to controls would be a curious result because H.M. scores low on working-memory measures whereas TTR and working-memory measures correlate positively for normal young and older adults (see Kemper & Sumner, 2001). Also problematic, Skotko et al. (p. 406) attribute H.M.'s low MLU, MCU, and LBC scores to his “inferior education,” a claim that suggests a flaw in their matching procedures. Moreover, contrary to Skotko et al., inferior education cannot in principle simultaneously explain both superior TTR scores for H.M. than for controls and (nonsignificantly) inferior MLU, MCU, and LBC scores for H.M. than for controls.