Intrasubject variability of cognitive and word retrieval performance across six sessions

Abstract

Florida State University, Tallahassee, FL 32306, USAA burgeoning interest appears to be developing in understandingthe nature and interactions between aphasia and nonlinguistic cogni-tive factors such as attention and working memory (Caspari, Parkin-son, LaPointe, & Katz, 1998; Erickson, Goldfinger, & LaPointe, 1996;Helm-Estabrooks, 2002; LaPointe & Erickson, 1991; Murray, 1999).Others (Hoffmann, 2001; Hochstenbach, Mulder, van Limbeek,Donders, & Schoonderwaldt, 1998) have noted the strong presence ofcognitive decline after stroke and commented on how marked slownessof information processing could impact linguistic performance.Some researchers (McNeil et al., 1991) have integrated disturbancesof cognitive processing into cognitive resource allocation models ofaphasia and suggested that the intrasubject variability of languageperformance seen in aphasia might be explained by fluctuating cog-nitive performance. Disturbances in word retrieval and lexical-se-mantic processing remains one of the hallmarks of aphasia andexplanations of variability across these domains also have been enig-matic (Moreno, Buchanan, & Van Orden, 2002; Nespoulous, Villiard,& Lecours, 1989). Despite the growing archive of research on cogni-tive-linguistic interactions, the sources of variability in aphasia remainelusive and unclearly understood. More elementally, the nature andrelationship of intrasubject variability on cognitive and word retrievaltasks across repeatedsessions of assessmentis relatively unstudied. Thepurpose of this study is twofold:1. To describe the variability of four individuals with aphasia on in-depth tasks of word retrieval and cognitive processing performanceacross six sessions of assessment.2. To explore the pattern of correlation and variability among wordretrieval and cognitive processing tasks across six sessions to deter-mine if changes in word retrieval across test sessions can be pre-dicted by changes in cognitive processing performance.MethodsParticipantsFour individuals with aphasia were recruited for this study andvolunteered to participate in six assessment sessions each. Our parti-cipants all suffered a single, left hemisphere thromboembolic CVA andwere mildly to moderately impaired in language as measured by theWestern Aphasia Battery and other standardized measures of languageimpairment. Participants were classified as having non-fluent, Broca’s-type aphasia. They ranged in age from 34 to 66 with a mean of 58.6.All were right handed pre-morbidly, were judged to be neurologicallystable beyond the acute epoch, and ranged in time post onset from 3months to several years. All participants were male.MeasuresSerialperformanceontheprinciplemeasuresincludedthefollowing:Cognitive measure: A computerized assessment protocol was se-lectedthatmeasuresthefollowingcognitiveskillsinthevisualmodality:• simple reaction time• choice reaction time• serial pattern matching• lexical discrimination• visual selective attention• response reversal with rapid visual scanning• abstract form discriminationThis measure has been standardized on a large group of non-braindamaged subjects and a variety of clinical populations (Miller, 1990;1995). The battery measures and records reaction times in millisecondsas well as accuracy across the 10 subtests of the full battery. Signaldetection analysis and accuracy corrected for false positives is alsorecorded and summarized.Word retrieval measure: The Test of Adolescent/Adult WordFinding (TAWF) (German, 1990). The complete TAWF batterymeasures naming nouns, verbs, sentence completion, naming to de-scription, and naming categories. Accuracy as well as latency of re-sponses is recorded.The measures were administered in counterbalanced order. The sixsessions for each participant were accomplished with approximatelyweekly sessions over a six-week period.ResultsFig. 1 presents the group means for the word retrieval scores (totalperformance for all TAWF subtests) as well as the group means for allof the CalCap cognitive processing subtests across the six sessions ofadministration. Description and analysis of performance across theindividual subtests and for individual subjects also will be presented.Spearman’s coefficient of correlation for word retrieval (groupTAWF total scores) and group CalCap reaction times was ).886(p = .028). Spearman’s r value for word retrieval and CalCap accuracywas ).08 (n.s.).