Monica S. Hough, Kun Yu, East Carolina University
Michael Cannito, University of Memphis
EFFECT OF AGE, WORKING MEMORY, AND
COMPREHENSION IN APHASIA
BACKGROUND INFORMATION: CONTEXT
Many aphasic individuals show deficits in
auditory comprehension.
 Examining the influence of context on
comprehension in aphasia has revealed that
linguistic context facilitates language
comprehension

(Cannito, et al., 1986; Cannito, et al., 1996; Hough, et al., 1989; Germani &
Pierce, 1992; Nicholas & Brookshire, 1983).
BACKGROUND INFORMATION: CONTEXT

Both predictive and nonpredictive linguistic context aid
auditory comprehension for aphasic individuals


specifically relative to reversible passive sentences
Context, particularly preceding predictive narratives,
limits possibility of upcoming events, making one
interpretation of a target sentence more plausible than
the other
(Brookshire, 1987; Germani & Pierce, 1992; Hough et al., 1989).

Even non-predictive linguistic context provides sufficient
redundant information to aid comprehension for aphasic
individuals.
BACKGROUND INFORMATION:
WORKING MEMORY (WM)




WM involves temporary storage and active manipulation of
information for complex cognitive processes, including
language comprehension
Baddeley’s multi-component WM system comprises a central
executive and two slave systems, phonological loop and
visuo-spatial sketchpad.
The phonological loop requires allocation of WM resources
Limitations in WM may lead to impairment in auditory
comprehension
still controversial and requires further investigation
(Baddeley, 1992, 1998, 2003; Caplan & Waters, 1999; Friedmann
& Gvion, 2003; Salthouse, 1994).

BACKGROUND INFORMATION: WM

Barrett, Tugade, and Engle (2004): individuals with low WM capacity
have difficulty with complex sentences




due to the inability to keep all necessary information active in WM.
individuals with low capacity need more time to respond to syntactically
ambiguous questions.
Caplan and Waters (1999): resources used for syntactic processing
in sentence comprehension may be separate from verbal WM
capacity.
Friedmann and Gvion (2003): two types of WM involved in sentence
processing;


syntactic WM for sentence comprehension
phonological WM for reactivation of word forms.
BACKGROUND INFORMATION: WM AND AGE
Advanced age also has been associated with
reduced WM functioning
 Wingfield et al. (1988): span test used to measure
WM capacity revealed remarkable age-related
differences between young group, older individuals
<70, and elderly individuals >70
 Dobbs and Rule (1989): aging has considerable
effect on processing aspect of WM, but lesser
effect on storage component.

(Byrne, 1998; Swanson, 1999; Wingfield, et al., 1988).
BACKGROUND INFORMATION: WM, AGE,
APHASIA

General consensus: many individuals with
aphasia have reduced WM capacity
(Caspari, et al., 1998; Francis, et al., 2003; Friedmann & Gvion, 2003;
Haarmann et al., 1997; Wright, et al., 2007; 2003).

The WM impairment appears to be one factor
contributing to comprehension deficits in
aphasia
(Caplan & Waters, 1997; Caspari et al, 1998; Davis, 2007; Hough,
Vogel, Cannito, & Pierce, 1997; Miyake et al., 1994).
BACKGROUND INFORMATION: WM, AGE,
APHASIA


Cannito et al. (1996) : limited WM resources led to reduced
comprehension of complex sentences for aphasic participants
Hough et al. (1997) examined performance of aphasic adults
presented with sentences in isolation and in contexts.



older subgroup of aphasic individuals demonstrated more accurate
performance on passive sentences and poor performance on active
sentences.
older aphasic individuals may have declined WM capacity; thus, tended
to identify last noun as agent of target sentences.
Caspari et al. (1998) : aphasic individuals with smaller WM capacity
performed more poorly than those with larger WM capacity in
comprehending complex sentences.
BACKGROUND INFORMATION:
WM, AGE, APHASIA
Investigating language performance variability in
aphasia relative to age, WM capacity, and
utilization of linguistic context to enhance auditory
comprehension may provide more insight into
aphasic patients’ ability to process information in
daily listening situations.
 However, few studies have examined interactive
effects of these variables and their influence on
use of linguistic context to aid comprehension in
aphasia.

PURPOSE OF CURRENT INVESTIGATION
The purposes of this study were:
1. to investigate influence of linguistic context on
auditory comprehension in aphasic individuals
2. to explore these influences considering effects
of age, WM, aphasia severity, and auditory
comprehension; and
3. to examine relationships among these
variables in adults with aphasia.

METHOD
PARTICIPANTS
16 adults with aphasia secondary to left
hemisphere CVA.
 divided into two groups based on age, with
eight participants > 60 and eight less < 60
 All participants passed a modified hearing
screening with at least 40 dB HL at 1000 Hz in
better ear.

Name
Young
1
2
3
4
5
6
7
8
Mean
SD
Range
Old
9
10
11
12
13
14
15
16
Mean
SD
Range
Gender
Age
Months Post-stoke
Education level
M
M
M
M
F
F
M
F
54
56
53
58
46
54
58
45
53
4.99
45-58
56
56
29
60
140
7
3
3
44.25
45.85
3-140
Graduate school
High school
College
Graduate school
High school
High school
High school
High school
M
M
M
M
F
F
F
F
74
73
61
63
63
69
84
86
71.63
9.53
61-86
91
12
57
197
59
15
115
6
69.00
64.77
6-197
Graduate school
High school
College
College
College
College
High school
5th grade
PARTICIPANTS


The Western Aphasia Battery-R (WAB-R) (Kertesz, 2006)
administered to all participants to determine presence and
severity via Aphasia Quotient (AQ)
Two subtests of Boston Diagnostic Aphasia Examination-III
(BDAE III) (Goodglass, Kaplan & Barresi, 2000), Oral
Commands and Complex Ideational Material, administered
to all aphasic individuals.


To be eligible for participation: individual had combined auditory
comprehension score on these subtests of 4 > 22
ensure presence of impairment as well as ability to perform
experimental tasks.
EXPERIMENTAL TASKS: WM CAPACITY





Listening span: measure each individual’s WM capacity.
modified version of Daneman and Carpenter’s (1980)
reading span task with modifications by Caspari et al.
(1998).
Participants asked to listen to sentence(s), remember
terminal word(s) for later recall, and answer questions
about sentences
Task included sentences at six levels with each level
including three trials
Terminal words were not related to sentences that they
were paired with.
Set1
Trial 1 The pilot flew the plane. Bread
Trial 2 The chef ruined the meal. Train
Trial 3 The phone rang again. Fork
Set 2
Trial1 Shelly ate another peach. Vest
Bill wrote a letter. Coat
Trial 2 He ordered steak and fries. Clock
He arrived too late. Queen
Trial 3 The thief escaped from prison. Stop
He didn't want any juice. Dish
EXPERIMENTAL TASKS: WM CAPACITY






At first level, three trials: each trial had one sentence and
one terminal word.
Participant was presented with an index card with one typed
sentence and one terminal word read orally.
Immediately after oral presentation, index card was turned
over to show three pictures, one corresponding to target
word.
Participant required to identify target picture by pointing.
This was considered one trial.
Participant had to select target pictures correctly in 2/3
trials to proceed to next level.
EXPERIMENTAL TASK: LANGUAGE
COMPREHENSION


Modified version of the Token Test (Caspari et al.,
1998): measure participants’ comprehension level and
investigate influence of WM on comprehension
Three levels of task: each level had ten commands
structured in increasing length and difficulty.



Comprehension task 1, only large tokens
Comprehension task 2, all twenty tokens with longer and
more difficult commands
Comprehension task 3, five tokens of different colors and
shapes
EXPERIMENTAL TASKS: CONTEXTUAL
INFLUENCES
consisted of 13 target sentences and 2 practice items in six
conditions:
 (1) reversible passive sentences in isolation
 (2) reversible active sentences in isolation
 (3) reversible passive sentences preceded by paragraphs predicting
specific subject-object relations of target sentence
 (4) reversible active sentences preceded by paragraphs predicting
specific subject-object relations of target sentence
 (5) reversible passive sentences preceded by paragraphs NOT
predicting specific subject-object relations of target sentence
 (6) reversible active sentences preceded by paragraphs NOT
predicting specific subject-object relations of target sentence
(Cannito et al., 1996; Hough et al., 1989; 1997)

Isolated context for active sentences
The nurse called the doctor.
Predictive context for active sentences
Both nurses and doctors work in a hospital. This hospital was overrun by patients. A
nurse began checking on the condition of a patient whose heart monitor was buzzing.
Suddenly, there was a frantic call through the ward. The nurse called the doctor.
Non-predictive context for active sentences
Both nurses and doctors work in a hospital. This hospital was overrun by patients. A
nurse began checking on the condition of a new patient. Suddenly, there was a frantic call
through the ward. The doctor called the nurse.
Isolated context for passive sentences
The doctor was called by the nurse.
Predictive context for passive sentences
Both nurses and doctors work in a hospital. This hospital was overrun by patients. A
nurse began checking on the condition of a patient whose heart monitor was buzzing.
Suddenly, there was a frantic call through the ward. The doctor was called by the nurse.
Non-predictive context for passive sentences
Both nurses and doctors work in a hospital. This hospital was overrun by patients. A
nurse began checking on the condition of a new patient. Suddenly, there was a frantic call
through the ward. The nurse was called by the doctor.
EXPERIMENTAL TASKS: CONTEXTUAL
INFLUENCES

For each stimulus item:




4 pictured response choices
consisting of black-and-white line drawings depicting two
possible subject-object relationships of target sentences,
subject-object relationship of related sentence, subjectobject relationship of unrelated sentence.
Contextual stimulus items presented auditorilly via
SuperLab 4.0 (Cedrus Corporation, 2006) with pictures
on Dell laptop
Participants were instructed to “show me what
happened” by choosing between pictures
RESULTS
AGE, WM, SEVERITY, COMPREHENSION
Descriptive statistics: means, standard
deviations, and ranges
 For Listening Span (WM), modified Token Test
(Auditory Comprehension), age, and WAB-R AQ
(Aphasia Severity)

Groups
Age
Aphasia severity
(WAB-R AQ)
(Max=100)
Working memory
(Max=21)
Auditory comprehension
(Max=29)
Mean
71.63
54.00
0.88
14.63
SD
9.53
32.38
0.44
12.21
Range
61-86
3.00-89.30
0-1.50
0-28
Mean
53.00
62.30
1.31
17.31
SD
4.99
26.04
0.92
7.79
Range
45-58
13.30-88.60
0.5-3.5
0-26
Mean
62.31
58.18
1.09
15.97
SD
12.10
28.70
0.74
9.99
Range
45-86
3.00-89.30
0-3.50
0-28
Old
Young
Combined
AGE, WM, SEVERITY, COMPREHENSION
Influence of explanatory variables upon one
another was examined via Pearson Product
Moment correlations.
 Effect of age as a continuous variable on WM,
aphasic severity, and comprehension: no
significant findings.

Aphasia
Severity
Comprehension
Working
memory
Auditory comprehension
Young
0.81**
Old
0.84***
Combined
0.83***
Working memory
Young
0.45
0.65*
Old
0.67*
0.66*
Combined
0.50**
0.57**
0.05
0.10
Age
Group
-0.16
AGE, WM, SEVERITY, COMPREHENSION




Relationships between WM and auditory comprehension,
WM and aphasia severity, and comprehension and severity:
linear regression lines calculated on scatter plots .
Comprehension and WM: slope of regression line was 7.74
(p=0.021); thus, comprehension increased as WM
increased.
WM and aphasia severity (WAB-R AQ): slope of regression
line was 0.01 (p=0.049); thus, WM increased with
decreased severity.
Comprehension and severity: slope of regression line was
0.28 (p<0.0001), indicating increasing comprehension with
decreasing aphasia severity.
Young
Old
Young
Old
Young
Old
EFFECT OF LINGUISTIC CONTEXT ON
COMPREHENSION

Influence of linguistic context on comprehension:




examined difference between scores for isolated active sentences and
active sentences preceded by predictive context and compared to same
differences for non-predictive contexts.
examined difference between scores for isolated passive sentences and
passive sentences preceded by predictive context and compared to
same differences for non-predictive contexts
Difference scores for active sentences: 8 participants demonstrated
comprehension improvement with predictive context s; 13
participants benefited from non-predictive contexts .
Difference scores for passive sentences: 12 participants benefitted
from predictive whereas 11 participants exhibited improvement for
non-predictive contexts.
EFFECT OF LINGUISTIC CONTEXT ON
COMPREHENSION


Paired sample t-tests conducted on contextual data: compare
performance between isolated active and passive sentences,
predictive active and passive context, and non-predictive active
and passive context; age as continuous variable.
Significant differences between:




isolated active and passive sentences (p<.05). Mean performance on
isolated active sentences was significantly higher than on isolated
passive sentences.
non-predictive active and passive sentences (p<05). Non-predictive
performance for active sentences was significantly higher than passive
sentences
no significant differences for predictive sentences..
For active sentences: mean performance was significantly higher with
predictive than non-predictive context (p<.05).
Iso1 active
–
Iso passive
Pre2 active
–
Pre passive
Non-pre3
active –
Non-pre
passive
Pre active –
Non-pre
active
Pre
passive–
Non-pre
passive
Mean
Std.
Deviatio
n
2.18750
3.03795
Paired Differences
Std.
95% Confidence
Error
Interval of the
Mean
Difference
Lower
Upper
.75949
.56869 3.80631
.25000
5.55578
1.38894 -2.71047 3.21047
.180
15
.860
3.62500
5.37742
1.34436
6.49043
2.696
15
.017*
-2.87500
4.31856
1.07964 -5.17620 -.57380
-2.663
15
.018*
.50000
5.34166
1.33542 -2.34637 3.34637
.374
15
.713
.75957
t
df
Sig.
(2-tailed)
2.880
15
.011*
EFFECT OF LINGUISTIC CONTEXT ON
COMPREHENSION



Relationships between explanatory variables and active/passive differences
for isolated, predictive, and non-predictive contexts examined via Pearson
Product Moment correlations
Age as a cpntinuous variable
Significant positive correlations between:




aphasia severity and differences for both isolated and non-predictive contexts
WM capacity and differences for isolated contexts.
No relationships between auditory comprehension and differences for any
contextual condition.
Relationships between explanatory variables and predictive/non-predictive
differences for active and passive sentences examined via Pearson
Product Moment correlations

no significant findings.
Active/passive
differences for
isolated
Active/passive
differences for
predictive
Active/passive
differences for
Non-predictive
Pre/Non
differences
for active
Pre/Non
differences
for passive
Aphasia
severity
0.463*
0.149
0.453*
-0.386
-0.011
Working
Memory
0.447*
0.316
0.337
0.009
0.019
0.192
0.228
-0.347
-0.251
Auditory
0.425
comprehension
DISCUSSION
DISCUSSION




Present study: investigated influence of linguistic context on auditory
comprehension in aphasic individuals considering effects of age, WM,
aphasia severity, and auditory comprehension.
For this group of aphasic individuals: no significant relationship observed
between age and WM.
As mentioned previously (Dobbs & Rule, 1989; Wingfield et al., 1988) found
large age differences in WM capacity between typical young and older
adults.
16 participants in current study:




1 young participant scored 3.5 for WM, considered high span.
All other participants scored 1.5 or below, considered low span.
Consider results with caution due to limited sample size.
Listening Span task was linguistically loaded; thus, participants may have
been at a disadvantage, with WM span underestimated.
DISCUSSION
No significant relationship between age and
severity of aphasia.
 Current findings in congruence with Obler, et al.
(1978): age had no impact on aphasia severity.
 No significant relationship between age and
auditory comprehension

DISCUSSION

Relationships between WM capacity and auditory
comprehension and WM capacity and aphasia severity:




auditory comprehension skills increased as WM capacity
increased
WM capacity increased with decreasing severity of aphasia.
Results consistent with Cannito et al. (1996): reduced WM
capacity leads to difficulty understanding sentences in
aphasic individuals.
Results also support Hough et al. (1997): limited WM
capacity had a negative impact on auditory comprehension
of sentences in aphasic adults.
DISCUSSION

Comparisons between conditions on the linguistic context task:



Aphasic individuals performed significantly better on isolated active than
passive sentences.
Consistent with previous reports: passive sentences are more difficult to
comprehend than active sentences for aphasic individuals (Berndt,
Mitchum, & Haendiges, 1996; Davis, 2007; Shewan & Canter, 1971).
All target sentences on context task were reversible.



Reversible sentences do not provide internal semantic constraints
aphasic adults had to rely on syntactic structure to comprehend sentences.
Consequently, participants’ auditory comprehension was negatively impacted
because word order in passive sentences is not linear and thematic role cannot
be assigned to first noun of sentence.
DISCUSSION





Aphasic adults performed significantly better on non-predictive than
predictive context for active sentences.
Active sentences were relatively easy to comprehend compared to
the passive sentences.
Thus, participants may have not required additional redundancy and
semantic support of the predictive contexts
Consequently, they may have ‘lost interest” when they heard target
sentences which provided “old” information consistent with
preceding predictive context.
In non-predictive context for active sentences, participants may have
been more linguistically engaged as they heard target sentences
which provided “new” information not alluded to in preceding
paragraphs.
DISCUSSION
However, non-predictive context may be less
facilitative than predictive context in
comprehension of more complex passive
sentences.
 This robust finding is consistent with previous
research and continues to require further
exploration relative to its use in language
treatment in aphasia.

Thank you for listening!
Questions???
[email protected]