Influence of Age and Executive Functions on Story Grammar Performance in Adults Michael S. Cannizzaro

1
Influence of Age and Executive Functions on Story
Grammar Performance in Adults Michael S.
Cannizzaro Carl A. Coelho University of
Connecticut, Storrs

• INTRODUCTION
• Previous investigations of age-related
  cognitive-linguistic decline have demonstrated
  the value of studying discourse ability (Chapman,
  Zientz, Weiner, Rosenberg, Frawley, Burns,
  2002 Glosser Desser, 1992 Makenzie, 2000
  Snow Douglas, 2000 Ulatowska, Freedman-Stern,
  Doyel, Macaluso-Haynes, North, 1983).
  Discourse can be conceptualized as the related,
  extended, and meaningful representation of
  communication across multiple language units (van
  Dijk Kintsch, 1983). Accurate production and
  comprehension of narrative discourse requires a
  complex interaction of linguistic, cognitive, and
  social abilities (Coelho, 1995 Ylvisaker,
  Szekeres, Feeney, 2001). The story narrative
  genre has frequently been studied because it is
  considered a stereotypical type of discourse due
  to its formal and predictable framework (Faylor
  Lemaire, 1993). Such narratives follow a
  cognitively based, measurable, structured set of
  rules called story grammar, which are not
  dictated by the specific content of the message
  conveyed (Coelho, 1998 Liles, Duffy, Merritt,
  Purcell, 1995 Mandler Johnson, 1977 Stein
  Glenn, 1979). Liles (1985) describes this
  underlying rule structure in narratives as a
  cognitive skeleton that creates character
  identities as well as the relational and causal
  elements that are fundamental to a well-formed
  story narrative. The production of discourse
  combines the use of language and cognitive
  organizational principles permitting
  investigation of the interaction of these
  abilities. Thus, discourse analyses are
  concerned with the entire discourse text and the
  relationship between its structure and function
  (i.e., the language itself and organization of
  the text) (McDonald, 1998). Indeed, discourse has
  been aptly described as one of the critical
  points of intersection between language and
  cognition as cognitive rule systems are employed
  to guide language production (Ylvisaker,
  Szekeres, Feeney, 2001, p. 756).
• It has been suggested that the episode structure
  of story grammar is one aspect of goal directed
  thinking and behavior referred to as managerial
  knowledge (Grafman, 1995 Grafman, 1999).
  According to Grafman (1995) managerial knowledge
  should be measurable in a variety of goal
  directed behaviors such as story grammar or
  executive functioning. Consistent with this
  notion are recent findings that have noted modest
  correlations between discourse production
  abilities and executive functions (Coelho, 2002
  Tucker Hanlon, 1998).
• Similarly, interest in age related changes
  across the life span has led to an increase in
  the number of investigations of the influence of
  cognitive performance and ageing on discourse
  (Chapman et al., 2002 Mackenzie, 2000).
  However, relationships between discourse and
  theoretically similar goal directed behavior,
  such as executive functioning, are not well
  established. The purpose of the present study
  was to investigate the influence of two factors,
  aging and executive functioning, on story grammar
  abilities in a large group of neurologically
  intact adults. Prior research suggests that age
  is related to declines in executive functioning
  (Delis, Kaplan, Kramer, 2001). It was
  hypothesized that performance on complex
  executive functions and narrative story structure
  would be similarly negatively correlated with
  advanced age, as both are goal conceptually
  related aspects of managerial knowledge.
• STUDY QUESTIONS
• What is the nature of the relationship between
  age and story grammar performance?
• Is story grammar performance related to measures
  of linguistic- and non-linguistic executive
  functions?
• Do EF measures and story grammar measures
  demonstrate a unified construct of Managerial
  knowledge?

METHODS Participants Forty-Six neurologically
intact adults (age 18-98, mean 56.78, sd 27.7)
were recruited for this study. All
participants were living independently in their
respective communities and reported negative
history of neurological and/or psychiatric
disorder, substance abuse, and language/learning
disability (Table 1).
Table 2 Story Grammar and demographic
correlations
Table 1 Age, EVIQ, and Years of Education
Table 3 Story grammar and linguistic-executive
function correlations














Discourse Analysis Procedures 1. Story
Retelling condition participants looked through
a picture story book and were instructed to tell
a story that would go along with the book.. 2.
Story Generation condition participants looked
at a single picture and instructed to tell a
story that would go along with that picture. In
both conditions, participants were instructed to
tell the story as if they were telling it to
someone who has never seen the book/ picture
before. No time limits were imposed in either
task. All stories were recorded, transcribed
verbatim, parsed into T-units, and analyzed for
story grammar content. Measures 1. Number of
Complete Episodes-an episode consisted of (a) an
initiating event that prompts a character to
formulate a goal-directed behavior, (b) an
action, and (c) a direct consequence marking
attainment or non-attainment of the goal. 2.
Proportion of T-units Within Episode
Structure-this measure was considered to be an
indication of participants ability to use story
grammar as an organizational plan for language.
3. Measures of Executive Function Seven
subtests of the Delis-Kaplan Executive Function
System (Delis, Kaplan, Kramer, 2001) were
administered to each participant. These
subtests were selected because they are known to
be sensitive to subtle differences in EF ability
and aging.
Table 4 Story grammar and non-linguistic-executi
ve function correlations
Table 5 Factor analysis - story grammar and
executive function measures
RESULTS Two measures of story narrative
performance, number of complete episodes in story
generation and proportion of T-units within
episode structure in story retelling, were
significantly negatively correlated with age, but
not participants level of education or estimated
verbal IQ (see Table 2). These two measures were
also significantly correlated with a number of
complex EF measures and participants motor speed
(see Tables 3 4) Hierarchical cluster
analysis based upon story grammar measures
identified three distinct clusters. A
Tukey-Kramer test for honestly significant
differences between groups (e.g., cluster mean
age Older mean 72.3 sd 17.6 Mid mean 61.7
sd 26.14 Young mean 49.5 sd 22.3 )
indicated the oldest age cluster was
significantly different from the youngest age
cluster. The intermediate age cluster was not
significantly different from either the younger
or older age clusters (see Figure 1).
Principal components analysis was performed on
the correlated discourse and age related
variables. Varimax rotation of variables yielded
two distinct factors (1) Output-Fluidity
accounting for 39 of the variance, and (2)
Organization-Efficiency accounting for 21 of the
variance. Greater than 60 of the variance
explained overall was accounted for by these two
factors. (see Table 5).
Figure 1 Story grammar cluster by age
DISCUSSION These findings are consistent with
previous investigations of cognitive-linguistic
changes in aging which have noted that discourse
changes are more pronounced in the older elderly,
the noted discourse changes are present even in
well educated elderly, age related changes are
more evident on complex tasks, and they occur
across a variety of discourse genres (Ulatowska,
Cannito, Hayashi, Fleming, 1985). Specifically,
decreased episode production in story generation
and inclusion of non-episodic information in
story retelling were notably different in the
older participants in this study. Story grammar
appears to be related equally to measures of
linguistic and non-linguistic executive function
and may provide a balanced ecologically valid
measure of executive ability. Additionally,
profiles of narrative performance (e.g., cluster
analysis) may provide a more appropriate
benchmark of elderly performance when assessing
and treating older adults with cognitive-communica
tive deficits.
Poster presented at the 33rd annual Clinical
Aphasiology Conference, May 2003. E-mail
michael.cannizzaro_at_uconn.edu