int. j. lang. comm. dis., 2002, vol. 37, no. 4, 459–474
Complex language functions and
subcortical mechanisms: evidence from
Huntington’s disease and patients with
non-thalamic subcortical lesions
Helen J. Chenery, David A. Copland and Bruce E. Murdoch
Department of Speech Pathology and Audiology, University of Queensland,
Queensland, Australia
(Received November 2001; accepted April 2002)
Abstract
The neuropathological changes associated with Huntington’s disease (HD) are
most marked in the head of the caudate nucleus and, to a lesser extent, in the
putamen and globus pallidus, suggesting that at least part of the language
impairments found in patients with HD may result from non-thalamic subcortical (NTS) pathology. The present study aimed to test the hypothesis that a
signature proŽ le of impaired language functions is found in patients who have
sustained damage to the non-thalamic subcortex, either focally induced or
resulting from neurodegenerative pathology. The language abilities of a group
of patients with Huntington’s disease (n 5 13) were compared with those of an
age- and education-matched group of patients with chronic NTS lesions following stroke (n 5 13) and a non-neurologically impaired control group (n 5 13).
The three groups were compared on language tasks that assessed both primary
and more complex language abilities. The primary language battery consisted
of The Western Aphasia Battery and The Boston Naming Test, whilst the more
complex cognitive–linguistic battery employed selected subtests from The Test
of Language Competence—Expanded, The Test of Word Knowledge and The
Word Test—Revised. On many of the tests of primary language function from
the Western Aphasia Battery, both the HD and NTS participants performed in
a similar manner to the control participants. The language performances of the
HD participants were signiŽ cantly more impaired ( p < 0.05 using modiŽ ed
Bonferroni adjustments) than the control group, however, on various lexico–
semantic tasks (e.g. the Boston Naming Test and providing deŽ nitions), on
both single-word and sentence-level generative tasks (e.g. category  uency and
formulating sentences), and on tasks which required interpretation of ambiguous, Ž gurative and inferential meaning. The diYculties that patients with HD
experienced with tasks assessing complex language abilities were strikingly
Address correspondence to: Helen J. Chenery, Center for Research in Language Processing and
Linguistics, Department of Speech Pathology and Audiology, University of Queensland, Queensland
4072, Australia; e-mail: [email protected]
International Journal of Language & Communication Disorders
ISSN 1368-282 2 print/ISSN 1460-6984 online © 2002 Royal College of Speech & Language Therapists
http://www.tandf.co.uk/journals
DOI: 10.1080/136828202100000773 0
460
H. J. Chenery et al.
similar, both qualitatively and quantitatively, to the language proŽ le produced
by NTS participants. The results provide evidence to suggest that a signature
language proŽ le is associated with damage to the non-thalamic subcortex
resulting from either focal neurological insult or a degenerative disease.
Keywords: Huntington’s disease, language.
Introduction
Huntington’s disease (HD) is an autosomal-dominant, neurodegenerative disorder
associated with neuropathological changes that are most marked in the head of the
caudate nucleus, and to a lesser extent the putamen and globus pallidus (Zakzanis
1998). Whilst several recent investigations into the neurocognitive features associated with HD have added considerably to a more accurate characterization of the
illness, the same cannot be said of the noted language deŽ cits associated with HD.
Recent language investigations have tended to focus on circumscribed aspects of
linguistic function, with few studies attempting comprehensively to proŽ le the
spared and disturbed language functions of people with HD. Additionally, given
the noted neuropathological focus in HD, namely the non-thalamic subcortex, it is
possible that the proŽ le of language abilities of people with HD may be similar to
that reported in patients with damage to the non-thalamic subcortex due to focal
cerebral insult. The present study sought to proŽ le the language abilities of patients
with HD on a comprehensive battery of language assessments and compare their
performances with a matched control group of aphasic patients whose cerebral
lesions were restricted to the non-thalamic subcortex.
HD is characterized, at a cognitive level, by impaired performances on tests of
delayed recall, measures of memory acquisition, cognitive  exibility and abstraction,
manual dexterity, attention/concentration, performance skill and verbal skill
(Zakzanis 1998). Investigations of the more speciŽ c cognitive deŽ cits that may
arise in HD have highlighted problems in (1) spatial working memory and visual
recognition memory and planning (Lawrence et al. 1998), (2) cognitive  exibility
and complex integration (Hanes et al. 1995), (3) attentional set-maintenance and in
the control processes that are assumed to prepare or to switch cognitive ‘sets’ in
order to perform one or another task (termed set-shifting) (Hanes et al. 1995,
Lawrence et al. 1998) and (4) problems with slower thought processes (termed
bradyphrenia) (Hanes et al. 1995). Behavioural manifestations of these impairments
are described in terms of a proclivity for impulsive behaviour (Hanes et al. 1995,
Butters et al. 1998) and perseveration of responses appropriate to contexts that have
not yet been negated (Lawrence et al. 1998).
Early studies of language functioning in subjects with HD reported a loss of
conversational initiative, visual confrontation-naming deŽ cits that were closely
related to visual misperception of the picture, and reduced syntactic structure of
spontaneous speech (Podoll et al. 1988). DiYculties in reading aloud and dysgraphia
were also prominent features of the HD performances but were suggested to be a
consequence of dysarthria and choreiform movement disorder. Podoll et al. (1988:
1475) concluded that ‘there are no primary language changes in HD, [but rather] a
variety of language impairments develop secondary to other neurological and neuropsychological changes’. Wallesch and Fehrenback (1988) went further to suggest
Language function in Huntington’s disease
461
that the syntactic deŽ cits observed in spontaneous speech and comprehension,
along with poorer naming scores, were the result of the HD subjects’ dementia.
These results contrasted with those from an earlier study reported by Caine et al.
(1986), who ensured that no HD subject who participated in their study was
pervasively demented. They reported a pattern of defective naming, impaired
repetition and decreased language output in written narratives by patients with HD
who were very early in the course of their illness, and suggested that a subtle but
‘deŽ nite language dysfunction’ ( p. 253) appears during the earliest stages of the
disease.
More recent studies have investigated discrete linguistic functions in people
with HD, particularly lexical–semantic functioning. In their study of word-list
generation using both initial letter and category cues, Barr and Brandt (1996) found
that subjects with more severe impairment (as determined by scores from the
QuantiŽ ed Neurological Examination; Folstein et al. 1983), exhibited greater impairment on semantic rather than phonemic generation tasks. The authors raised the
possibility that a disturbance in the organization of semantic memory may occur
early in the course of HD but that increasing disease severity and associated greater
generalized retrieval diYculties may obscure this speciŽ c semantic impairment.
Con icting Ž ndings by Hodges et al. (1991) were reported on a confrontationnaming task, however, with no evidence found to support a semantic basis to the
naming deŽ cit. Rather, visuoperceptual and visuospatial deŽ cits were thought to
contribute to the higher proportion of visually based errors by the HD subjects.
Also of interest in the literature is whether the language changes associated with
HD form a pattern of deŽ cits that is consistent with subcortical pathology. Whilst
the striatum is most aVected by neuronal death in HD, other subcortical sites, as
well as cerebral cortex, have also been implicated (Selemon et al. 1998). As stated
previously, the neuropathological changes are most marked in the head of the
caudate nucleus and, to a lesser extent, the putamen and globus pallidus (Zakzanis
1998), suggesting that the language changes found in people with HD may share
some features with those observed in people with lesions to the striatum resulting
from subcortical stroke. Converging evidence from focal lesion studies is highlighting a potential role of the striatum in aspects of language processing, although
this issue is still far from clear (see Special Issue ‘Brain and Language’, vol. 58,
1997). A number of studies have reported a variety of language deŽ cits in patients
who have sustained striatocapsular lesions as the result of cerebrovascular accident,
including impairments in, for example, auditory comprehension (D’Esposito and
Alexander 1995), lexical–semantic processing (Vallar et al. 1988) and complex
language tasks (Copland et al. 2000). The core language proŽ le of patients with
striatocapsular infarction was described by Mega and Alexander (1994) as including
deŽ cits in executive language functions such as word  uency, sentence generation
and discourse compared with relatively spared responsive language, namely
comprehension, repetition and, in some milder cases, naming. It may be possible
to progress the study of language impairment in HD by comparing their language
test scores with a group of patients who have sustained discrete non-thalamic
subcortical (NTS) lesions as the result of cerebrovascular accident. This comparison
is undertaken in the present study.
Thus, the aims of the present study are to proŽ le the language abilities of a
group of patients with HD on a comprehensive battery of standardized language
assessments that are sensitive to a wide range of language functions and compare
462
H. J. Chenery et al.
the performances of the HD group with a control group matched for age, sex and
education. It is predicted the participants with HD would perform signiŽ cantly
more poorly than the control group, particularly on those tasks assessing lexical–
semantic and syntactic function. Additionally, the possible in uence of subcortical
language-processing mechanisms will be investigated by including a further comparison group of subjects who have sustained NTS damage due to cerebrovascular
accident. It is hypothesized that people with HD will evidence a language impairment
that is qualitatively similar to that exhibited by a group of people with focal
cerebrovascular accident aVecting the non-thalamic subcortex, suggesting a common
subcortical basis for these patterns of language impairment.
Methods
Subjects
Thirteen patients (Ž ve male, eight female) with HD participated in the study. They
averaged 52 (SD 10.60) years of age and 11.38 (2.43) years of education. The
patients had been diagnosed by a neurologist based on a positive family history for
the disease and the presence of involuntary choreiform movements. The mean
score for the HD group on the Dementia Rating Scale ( DRS; Mattis 1988) was
117.92 (SD 7.66). The DRS provides scores on Ž ve parameters: attention, conceptualization (verbal and non-verbal ), initiation/perseveration (verbal and motor),
memory (verbal and non-verbal ), and construction. These scores are combined to
obtain a general measure of cognitive status with a maximum of 144. In addition,
a general rating of functional ability was performed using the levels of disability
scale described by Shoulson and Fahn (1979). A rating of 1 on the scale represents
minimal disability with a 5 rating representing severe disability. The mean rating
for the group was 2.85 (SD 0.99). All patients were recruited through the local
branch of the Huntington’s Disease Association and were attending weekly support
meetings at the centre.
Two control groups were recruited to participate in this study. The Ž rst control
group comprised 13 healthy control subjects recruited from the community who
were matched to the HD subjects in terms of age (mean 56.54 years, SD 10.18),
sex and years of education (mean 10.69 years, SD 2.72). Exclusion criteria for the
control subjects included: (1) a previous history of neurological disease such as
stroke and progressive dementia, (2) a concurrent psychiatric illness, (3) a history
of developmental learning or language disability and (4) a history of hearing
impairment.
A second control group comprised 13 patients with computed tomographic
(CT)- or magnetic resonance imaging (MRI)-conŽ rmed lesions conŽ ned to subcortical regions, excluding the thalamus. These 13 participants are a subgroup of
participants previously reported in Copland et al. (2000). All patients had suVered
a single left cerebrovascular accident. Subjects were excluded if there was a positive
history of head trauma, dementia, brain tumour, cerebral abscess or alcoholism. All
subjects were monolingual in English and had no reported visual and/or hearing
abnormality. Testing was carried out at least 5 months post-onset in all subjects.
The mean (SD) age of the NTS group was 59.46 (16.04) years and their mean
educational level was 11.0 (2.35) years. Full demographic and neuroradiological
information for each subject with an NTS lesion is shown in table 1.
Language function in Huntington’s disease
Table 1.
463
Demographic and neuroradiological information for patients with non-thalamic
subcortical lesions
Age
Education
Time from stroke
Case (years) Sex (years) Aetiology to scan (days)
1
2
3
4
5
6
7
8
9
10
11
12
13
77
91
49
81
46
40
76
52
52
47
52
52
58
F
F
F
F
F
F
M
M
M
M
F
F
F
8
10
10
10
15
10
10
10
10
15
10
10
15
I
I
I
I
I
I
H
H
I
H
H
H
I
1
> 365
1
4
64
> 365
48
37
1
43
4
3
3
Lesion site
IC
GP
IC
PVWM, EC, BG
IC, BG, PVWM
PVWM, CS, LN, IC, EC
BG
BG
IC, HCN, LM
P, IC
CN, IC
CN, IC
DWM
Months
post-stroke
37
84
9
5
51
13
12
6
32
49
13
71
24
I, infarct; H, haemorrhage; IC, internal capsule; GP, globus pallidus; PVWM, perventricular white
matter; EC, external capsule; BG, basal ganglia; CS, centrum semiovale; LN, lentiform nucleus; HCN,
head of the caudate nucleus; P, putamen; CN, caudate nucleus; DWM, deep white matter.
Materials
A comprehensive language test protocol was constructed that served to assess a
wide range of language abilities, including more complex linguistic functions. The
Ž rst component of the battery (referred to as Battery 1) was concerned with primary
language processes, as assessed by the Western Aphasia Battery (WAB; Kertesz
1982) and the Boston Naming Test (BNT; Kaplan et al. 1983). The WAB was
selected as it tests a variety of language processes whilst remaining sensitive to both
severe and milder impairments. The BNT was included in the battery as it provides
a measure of word Ž nding ability across a range of word frequencies.
The second component of the assessment battery (Battery 2) comprised a range
of assessment tasks that were more sensitive to complex and demanding linguistic
processes. The language tasks selected were the Test of Language Competence—
Expanded Edition (TLC-E; Wiig and Secord 1989), and selected subtests from the
Test of Word Knowledge (TOWK; Wiig and Secord 1992) and The Word Test—
Revised (TWT-R; Huisingh et al. 1990).
The TLC-E is designed to probe divergent production, cognitive–linguistic
 exibility and planning for production. It consists of tasks that require problem
solving, planning and decision-making as well as providing alternative solutions
or responses for the same linguistic input. Four subtests were included, a full
description of which is provided in the appendix.
The TOWK assesses the ability to recognize and express critical semantic
features of the lexicon through tasks involving the selection of appropriate words
and the elicitation of word deŽ nitions. Four subtests were included, which are once
again more fully described in the appendix. Only one subtest of the TWT-R was
included to assess the capacity for verbal reasoning and the ability to recognize
critical semantic features of vocabulary.
464
H. J. Chenery et al.
Procedure
All administration and scoring of the assessments were carried out in accordance
with the directions as laid down in the respective test manuals. Test administration
took place under standard conditions, in a quiet environment. Testing was usually
completed over two to four sessions, with the proviso that testing was discontinued
if a subject showed any signs of fatigue. The order in which the various tests were
administered was randomized for each participant.
Results
Test Battery 1
Seven of the 11 subtests from Battery 1 showed marked ceiling eVects where one
or two participant groups attained a perfect score. To analyse group performance
on these subtests, participant scores were assigned a nominal score of either 1
(attained perfect score) or 2 (attained less than a perfect score). These frequency
tables were then analysed using Fisher’s Exact test with all tests having 1 d.f. To
account for the multiplicity of comparisons and the possibility of an in ated Type
I error, a more stringent a level of p < 0.01 was adopted.
The raw data for the remaining four subtests were initially examined for
normality, homogeneity of variance and linearity. To test for homogeneity of
variance, the procedure described by Tabachnick and Fidell (1995) using calculations
of Fmax in conjunction with sample size ratios was used. The Fmax ratio was
consistently < 10 indicating acceptable homogeneity of variance. The data contained
in this subset of tests was negatively skewed, however, and violated the assumption
of normality, so Kruskall–Wallis non-parametric analyses were performed. To correct
for the multiplicity of comparisons across the four subtests, a more stringent a
level was set, using a modiŽ ed Bonferroni procedure ( Jaccard and Wan 1996). For
comparisons across all four subtests, p < 0.0125 to achieve signiŽ cance, for three
subtests p < 0.0167, for two subtests p < 0.025, etc.
As a Ž rst step in proŽ ling the language performances of the participants with
HD, the individual scores from Test Battery 1 (presented as Scaled Scores from
the WAB) in descending order of cognitive severity as measured by the DRS are
shown in table 2. The means and SDs of subtest scores from Test Battery 1 as a
function of group (either HD, NTS or control) are shown in table 3. Using modiŽ ed
Bonferroni adjustment of a levels, only two subtests, namely Word Fluency and the
BNT, were signiŽ cantly diVerent across the three groups. Mann–Whitney U-tests
with modiŽ ed a levels were performed on the subtests of Word Fluency and the
BNT to determine which groups performed signiŽ cantly more poorly than the
other groups. Both the participants with HD and NTS lesions obtained signiŽ cantly
lower scores than the control subjects on the BNT (U 5 13.00, p < 0.001 and U 5
41.50, p 5 0.02, respectively) and Word Fluency (U 5 2.00, p < 0.001 and U 5 25.00,
p 5 0.002, respectively). The NTS group did not diVer from the HD group on either
Word Fluency or the BNT. For the subtest of Fluency, cross-tabs statistics using
Fisher’s Exact test with 1 d.f. showed that both the HD and NTS groups diVered
signiŽ cantly ( p < 0.01) from the control group, but not from each other. Only the
HD group diVered signiŽ cantly from the control subjects on the subtest of Sequential
Commands.
Western Aphasia Battery scaled scores and subtest scores for individuals with Huntington’s disease
128
127
125
123
122
121
119
118
115
115
109
108
103
45
31
29
51
46
37
47
46
48
52
34
37
53
94.4
98.2
95.6
96.4
95.6
93.4
97.4
94.6
90.2
97.4
92.2
93.8
97
19
20
19
19
19.5
18
20
19
18
20
18
19
20
10
9.8
9.8
9.9
9.3
9.8
9.7
10
10
10
9.1
9.7
9.8
9.7
9.8
9.9
10
9.7
9.8
9.9
9.6
8.7
9.7
10
9.6
10
8.5
9.5
9.1
9.3
9.3
9.1
9.1
8.7
8.4
9
9
8.6
8.7
38
40
23
34
32
38
18
34
33
40
39
31
40
50
47
21
49
35
44
15
33
29
48
37
31
54
36
35
15
36
35
38
12
40
37
39
34
32
41
23
24
5
26
28
18
15
14
15
28
25
15
26
18
17
6
14
19
19
7
13
6
30
22
8
29
23
30
20
24
26
28
11
23
15
19
22
20
28
69
58
47
39
48
60
63
69
40
62
59
49
65
30
32
12
31
18
29
12
24
17
28
23
24
34
12
13
2
13
12
14
9
8
13
15
13
11
14
DRS, Dementia Rating Scale; BNT, boston Naming Test; WAB AQ, Western Aphasia Battery Aphasia Quotient; SS, spontaneous speech; Com, comprehension;
Rep, repetition; Nam, Naming.
1
2
3
4
5
6
7
8
9
10
11
12
13
DRS
BNT
WAB
SS
Com
Rep
Nam
Word
Word
Multiple
Ambiguous
Oral
Figurative
Semantic
Case score/144 score/60 AQ/100 total/20 total/10 total/10 total/10 opposites/42 deŽ nitions/64 Synonyms/42 contexts/32 sentences/39 Inferences/36 expression/78 language/36 absurdities/15
Table 2.
Language function in Huntington’s disease
465
466
Table 3.
H. J. Chenery et al.
Comparative performance among Huntington’s disease, non-thalamic subcortical
and control subject on subtests from Battery 1
Huntington’s
disease group
(n 5 13)
Non-thalamic
group
(n 5 13)
Subtest
Mean
SD
Mean
SD
Mean
SD
x
Repetition (0–100)
Word  uency (0–20)
Boston Naming Test (0–60)
Auditory word recognition
(0–60)
97.23
10.38b
42.77b
59.31
3.39
2.81
8.17
1.55
98.85
12.77b
49.69b
59.85
1.68
4.02
5.07
0.38
99.31
17.92a
54.23a
59.92
1.25
2.43
3.39
0.28
6.23
20.21
14.92
1.40
Control group
(n 5 13)
2
CN/
HD
Information content (0–10)
Fluency (0–10)
Yes/no questions (0–60)
Sequential commands (0–80)
Object naming (0–60)
Sentence completion (0–10)
Responsive naming (0–10)
9.62
9.50b
59.69
76.69b
59.38
9.69
10.00
0.51
0.50
0.85
4.75
1.19
0.75
0.00
9.77
9.46b
59.77
79.00a
59.46
10.00
9.85
0.60
0.52
0.83
2.52
1.05
0.00
0.55
10.00
10.00a
60.00
80.00a
60.00
10.00
10.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.02
0.003x
0.24
0.007x
0.11
0.24
0.5
p
0.044
< 0.001*
0.001*
0.497
CN/ HD/
NTS NTS
0.24
0.005x
0.5
0.24
0.22
0.5
0.5
0.38
1.0
1.0
0.20
0.5
0.48
1.0
Numbers in parentheses denote the possible minimum and maximum scores obtainable on each subtest.
CN, control subjects; HD, Huntington’s disease; NTS, non-thalamic subcortical.
*SigniŽ cant at p < 0.05 with modiŽ ed Bonferroni adjustment. For comparisons on all four subtests,
p < 0.0125 to achieve signiŽ cance, for three subtests p <0.0167, for two subtests p <0.025 and so on
until p <0.05 for all four subtests. Thus, the Ž rst two tests reached the required level of signiŽ cance
at p <0.05.
x SigniŽ cant at p < 0.01.
Superscript letters indicate signiŽ cant ( p <0.01) diVerences between groups, i.e. where two groups are
identiŽ ed with the same letter, there were no signiŽ cant diVerences between them, but where they are
identiŽ ed with diVerent letters, their scores were signiŽ cantly diVerent.
Test Battery 2
Once again, individual scores from the HD participants obtained for the various
subtests from Test Battery 2 are shown in table 2 as a function of increasing
cognitive impairment (as measured by the DRS). Preliminary data screening of the
subtests in Battery 2 revealed acceptable homogeneity of variance levels, but again
a negative skewness in the data. Non-parametric Kruskall–Wallis tests were used to
analyse the diVerences among the three groups. Seven of the nine subtests from
Battery 2 were signiŽ cantly diVerent using modiŽ ed Bonferroni a levels (for comparisons across all nine subtests, p < 0.0055 to achieve signiŽ cance, for eight subtests
p < 0.0062, for seven subtests p < 0.007, etc.). These seven subtests were Ambiguous
Sentences, Making Inferences, Oral Expression, Figurative Language, Word
DeŽ nitions, Multiple Contexts and Semantic Absurdities (table 4). Overall, the
subjects with HD performed poorly, with Mann–Whitney comparisons revealing
signiŽ cantly ( p < 0.001) poorer performances than the control subjects on the seven
subtests from Battery 2 (table 4). The NTS subjects were also signiŽ cantly more
impaired than the control subjects on Ž ve of the Battery 2 subtests, namely
Ambiguous Sentences ( p < 0.001), Figurative Language ( p 5 0.001), Multiple
Language function in Huntington’s disease
Table 4.
467
Comparative performance among Huntington’s disease, non-thalamic subcortical
and control subjects on subtests from Battery 2
Subtest
TLC-E
Ambiguous sentences (0–39)
Making inferences (0–36)
Oral expression (0–78)
Figurative language (0–36)
TOWK
Word opposites (0–42)
Word deŽ nitions (0–64)
Synonyms (0–42)
Multiple contexts (0–32)
TWT-R
Semantic absurdities (0–15)
Huntington’s
disease group
(n 5 13)
Non-thalamic
group
(n 5 13)
Mean
Mean
SD
SD
Control group
(n 5 13)
Mean
SD
x
2
p
16.00b 8.07
22.23b 5.36
51.15b 16.58
24.15b 7.44
23.00b 8.82
27.31ab 5.04
59.46b 18.08
29.85b 5.37
35.62a
31.00a
75.38a
34.69a
3.25
4.76
2.33
1.60
24.38 <0.001*
13.42
0.001*
20.92 <0.001*
20.45 <0.001*
33.85
6.80
37.92b 11.98
33.08
9.05
20.15b 7.01
34.92
6.92
39.77b 10.65
35.23
5.83
23.38b 6.90
39.92
53.85a
40.31
30.38a
1.55
4.16
1.38
1.26
9.44
0.009
18.03 <0.001*
12.39
0.002
21.21 <0.001*
11.46b
14.15a
14.77a
0.44
18.65 <0.001*
3.45
1.14
Numbers in parentheses denote the possible minimum and maximum scores obtainable on each subtest.
*SigniŽ cant at p < 0.05 with modiŽ ed Bonferroni adjustment. For comparisons on all nine subtests,
p < 0.0055 to achieve signiŽ cance, for eight subtests p <0.0062, for seven subtests p <0.007, for six
subtests p < 0.008 and so on until p <0.05 for all nine subtests. Thus, the Ž rst seven tests reached the
required level of signiŽ cance at p <0.05.
Superscript letters indicate signiŽ cant ( p <0.05) diVerences between groups, i.e. where two groups are
identiŽ ed with the same letter, there were no signiŽ cant diVerences between them, but where they are
identiŽ ed with diVerent letters, their scores were signiŽ cantly diVerent.
Contexts ( p < 0.001), Oral Expression ( p 5 0.002) and Word DeŽ nitions ( p < 0.001).
Interestingly, the HD subjects and the NTS subjects performed in a very similar
manner. SpeciŽ cally, the results of the Mann–Whitney U-tests showed that the
subjects with HD were signiŽ cantly more impaired than NTS subjects ( p 5 0.002)
only on the subtest of Semantic Absurdities (table 4).
Discussion
The results of the study have demonstrated that people with HD perform signiŽ cantly more poorly on complex language tasks that assess a variety of cognitive–
linguistic abilities than a group of matched, control subjects. SpeciŽ cally, the HD
patients experienced diYculty on more demanding language tasks that involved
lexico–semantic manipulation, interpreting indeterminacy of meaning and propositional language. Their pattern of language impairment was strikingly similar to that
found in a matched group of patients with NTS lesions, suggesting that striatal
damage may underlie the language deŽ cits associated with HD.
The performance of patients with HD on primary language tasks such as the
WAB was mostly preserved and once again was similar to that found in an NTS
lesion control group. Previous reports of spared language performance on primary
language tasks in both HD (Podoll et al. 1988, Wallesch and Fehrenbach 1988) and
NTS patients (Godefroy et al. 1994, Mega and Alexander 1994) have been extended
by the results of the present study which included a more complex cognitive–
linguistic assessment battery. While the performance of both the HD and NTS
468
H. J. Chenery et al.
subjects on most subtests of the WAB gives the impression that the linguistic
system is generally intact, extended testing revealed signiŽ cant impairments on more
complex language tasks.
Evidence from the present study suggests that the language deŽ cit in patients
with HD is related to the speciŽ c pattern of striatal degeneration associated with
this illness. SpeciŽ cally, the pattern of performance of the subjects with HD was
strikingly similar to a group of age- and education-matched subjects with focal NTS
lesions, lending support for the proposal that a signature proŽ le of impaired language
functions may be found in patients with NTS damage, either focally induced or
resulting from degenerative pathology. This is not to suggest that the language
impairments in HD patients are not related to speciŽ c neurocognitive dysfunction
associated with striatal damage such as deŽ cits in set shifting or negation of contexts.
These speciŽ c hypotheses require further investigation. Rather, the current Ž ndings
highlighted deŽ cit performance of both the HD and NTS subjects on tasks assessing
(1) lexico–semantic processing, (2) generative language and (3) the interpretation
of sentential content, and suggest that the pattern of language performance in
patients with HD or NTS stroke may result from their common subcortical
pathology.
Lexico–semantic processing
A number of subtests in the present study assessed lexico–semantic processing and
people with HD performed signiŽ cantly more poorly than age-matched controls
on two of these subtests, namely the BNT and Word DeŽ nitions. The group of
patients with NTS lesions showed a similar pattern of performance, with no
signiŽ cant diVerences between the two groups on these two measures of lexico–
semantic processing.
DeŽ cits in confrontation naming have been frequently reported in previous
studies of language function in HD. Podell et al. (1988) reported a high proportion
of errors that were semantically associated to the target whilst noting that a further
category of purely visually related errors was also prevalent. Hodges et al. (1991)
also reported both visually based errors (that were produced signiŽ cantly more
often than a matched control group) and a high proportion of semantic errors (that
was similar to the proportion made by control subjects). Whilst a detailed analysis
of naming errors was beyond the scope of the present study, we noted that our
HD patients also tended to produce a higher proportion of semantically related
errors (e.g. ‘swordŽ sh’ for the target sea horse, ‘eskimos’ house’ for the target igloo,
‘thing I put my clothes on’ for hangar and ‘scary face’ for mask). To a lesser extent,
we noted visually related errors such as ‘didgeridoo’ for the target asparagus, ‘bookshelf ’ for abacus, ‘looks like a candle in a church’ for door knocker, ‘that’s a cup’ for
mask, and ‘olden-day clock’ for protractor. Like Podell et al. (1988) and Hodges et al.
(1991), a proportion of the patients’ errors (e.g. ‘sugar cane’ for asparagus) could
also reasonably be described as being both semantically and visually related (or
what Hodges et al. call ambiguous visual/semantic category errors), although it is
likely that a semantic basis underlies these errors as items from a similar semantic
category (e.g. things we eat) are also likely to share many structural characteristics.
Continued debate exists within the literature about the likely basis for these
naming errors. The presence of visually related errors in our HD patient group,
Language function in Huntington’s disease
469
and as reported in previous studies, may be linked to the well-established visuoperceptual and visuospatial deŽ cits of HD patients (Butters et al. 1978, Brouwers et al.
1984). In our study, however, semantically related errors did appear to dominate
the error types, a Ž nding similar to that reported by Frank et al. (1996) who found
no visual misperceptions by their HD group on a test of visual recognition. Rather,
Frank et al. suggested that HD is associated with a restricted access to and use of
semantic features for concept identiŽ cation, a proposal that oVers a parsimonious
explanation of the predominance of semantically related errors in our HD group.
The confrontation-naming deŽ cit of patients with HD was quantitatively similar
to the patients with NTS lesions, with their scores on the BNT not diVering
signiŽ cantly. At a qualitative level, the similarity between the HD and NTS subjects’
predominance of semantically related errors was noted. Both semantic paraphasias
(e.g. ‘bouquet’ for wreath and ‘checkers’ for dominoes) and deŽ nitional errors (e.g.
‘you make a circle with it’ for compass) were frequently noted in the sample of NTS
naming errors. Wallesch and Papagno (1988) suggested that lexical–semantic processing is commonly aVected in patients with NTS lesions, and the present results
support previous studies that report persisting naming deŽ cits and the presence of
semantic paraphasias following dominant striatocapsular lesions (Robin and
Schienberg 1990, Kennedy and Murdoch 1993, Mega and Alexander 1994). The
similarity between the BNT naming scores for the HD and NTS group suggests
that striatal degeneration may be the likely locus for this deŽ cit and conŽ rm an
early report by Caine et al. (1986), who reported the presence of a deŽ nite language
dysfunction (including defective naming) in non-demented patients with HD who
were early in the course of their disease.
The Multiple Contexts subtest, which assesses the ability to manipulate the
lexical–semantic system, was also performed poorly by subjects in the HD and the
NTS groups. The Multiple Contexts task required subjects to provide deŽ nitions
of homophonic words, that varied both in terms of frequency and concreteness/
abstractness (e.g. bat is both a higher frequency and concrete concept whereas neutral
is both lower frequency and abstract). The deŽ nitions produced by the HD subjects
mostly involved the subjects correctly providing one deŽ nition but then being
unable to shift set to provide an alternative meaning. For example, one HD subject
gave a meaning of the homophone mole as ‘small animal’ but responded ‘I can’t
think of another one’ when pressed for a second meaning. Whilst the majority of
the errors produced by the HD subjects were of this type, occasionally the HD
subjects perseverated on the Ž rst meaning (e.g. letter was deŽ ned as ‘something you
put in the mail’ and ‘piece of paper you write on’ or manual as ‘manual car’ and
‘drive manually’).
This pattern of response Ž nds a possible cognitive correlate in the noted
diYculties that patients with HD experience in set shifting (Hanes et al. 1995,
Butters et al. 1998, Lawrence et al. 1998) and in cognitive  exibility (Hanes et al.
1995). The type of perseverative behaviour found on this subtest has been reported
previously in HD (Lawrence et al. 1998) and was accounted for by a preferential
loss of the indirect striatal pathway that mediates negation of contexts, the behavioural eVects of which would be perseveration of responses appropriate to contexts
that have not yet been negated. Indeed the predominance of ‘no responses’ by our
HD patients on the second deŽ nition could also be due to diYculty in negating
the context that has been established for the Ž rst response, thereby impeding the
provision of another context-related meaning of the ambiguous word.
470
H. J. Chenery et al.
Generative language
DeŽ cits in word  uency are a frequently reported feature of the language proŽ le
of both HD (Caine et al. 1986, Rosser and Hodges 1994, Barr and Brandt 1996,
Zakzanis 1998) and NTS patients (Alexander et al. 1987, Vallar et al. 1988, Mega
and Alexander 1994). The results of these previous studies were conŽ rmed in the
present investigation with both HD and NTS subjects showing impaired performance on a category  uency task. Using tasks of letter and category (both animate
and inanimate)  uency generation, Rosser and Hodges (1994) found that their HD
patients gave signiŽ cantly fewer items than controls in both generation tasks. Barr
and Brandt (1996) also found reduced category  uency in their sample of HD
subjects, as did the present study, combined with signiŽ cantly reduced letter  uency.
The Oral Expression subtest of the TLC-E was particularly diYcult for both
the NTS and HD individuals. This subtest required subjects to incorporate three
target words into a self-generated sentence, which was to relate to a pictured
situation. On occasions, the HD subjects gave grammatical sentences that contained
only one of the three words. For example, instructed to use the target words without
diYcult wrong in a sentence that was to relate to a track meet, one HD subject
responded with ‘The race was too diYcult’. When the HD subjects attempted to
construct a sentence including more of the target words (e.g. using before rather after
in a sentence relating to a movie theatre), the sentences frequently were ungrammatical, e.g. ‘Before which would you rather? After too.’ Indeed, in many of the earlier
studies of language in HD, deŽ cits in syntactic complexity of spontaneous speech
(Wallesch and Fehrenbach 1988) or observations that spontaneous speech was
typically reduced to short, simple sentence constructions (Podoll et al. 1988) were
noted.
In the present study, participants with NTS focal lesions also demonstrated
diYculty on the Oral Expression subtest. Several of the NTS subjects produced
sentences that were syntactically or semantically anomalous, that were incomplete,
or that failed to include all three words in the recreated sentence. Mega and
Alexander (1994) also found that subjects with striatocapsular lesions exhibited
deŽ cits (manifested as increased response latencies, perseverations, and occasional
bizarre content) when asked to produce a sentence containing a speciŽ ed verb.
More detailed testing and error analyses are required to determine the precise basis
of these sentence production diYculties and their underlying similarities across the
HD and NTS groups.
Interpretation of sentential-level meaning
Four subtests were thought to assess the abilities of subjects to interpret complex
sentences. The Ambiguous Sentences task required subjects to provide two meanings
of an ambiguous sentence. The HD subjects once again were mostly able to provide
at least one deŽ nition, but in a similar way to the Multiple Contexts task, were
unable to then negate that meaning and provide a second interpretation.
Occasionally the patients gave two similar meanings for the sentence or simply
paraphrased the sentence without giving a precise interpretation (e.g. for the sentence, Can you believe Mary wanted to run as well as me?, one HD subject responded
with ‘that’s competition’). The NTS subjects also experienced diYculty with this
subtest, frequently giving only one meaning of the sentence, accompanied by deŽ cits
Language function in Huntington’s disease
471
in retrieving or processing another meaning of the sentence. NTS subjects also
occasionally provided two responses that were related to the same meaning of the
ambiguity. DiYculties at this level of indeterminacy of meaning may once again
re ect an impairment in negating information brought to consciousness and then
perceiving an alternative, a process thought to be mediated by the indirect pathway
of the striatum (Lawrence et al. 1998).
The Figurative Language subtest also requires an appreciation of the contextual
constraints imposed by a pictured stimulus prior to the subject explaining a verbally
presented metaphor. Occasionally, the response was related to the required interpretation (e.g. when one HD subject was asked to provide an explanation for the
metaphor It’s hard to zero in on his ideas in the context of ‘two students talking about
a teacher’, she responded ‘They don’t really like the teacher’). It appears that HD
subjects may have gained only a literal interpretation of the metaphor and could
not link this interpretation to the context in a meaningful way. Similarly, patients
with NTS lesions also performed signiŽ cantly more poorly on this subtest than
matched controls, a Ž nding that is consistent with Wallesch et al.’s (1983) suggestion
that patients with basal ganglia lesions are unable to adequately explain idiomatic
expressions.
DeŽ cits in interpreting sentential or paragraph information were also noted in
the HD subject group on the Making Inferences Subtest. The understanding of the
short paragraphs in this task requires the synthesis of both fundamental linguistic
processes and other cognitive abilities such as constructing inferences about
information that may not be directly stated in the text. In their assessment of
discourse comprehension using measures of both salience and directness, Murray
and Stout (1999) found that individuals with HD (as well as people with Parkinson’s
disease) experienced signiŽ cant diYculty in processing detailed or implied
information. Murray and Stout (1999) suggested that diYculty in integrating story
information with contextual or prior knowledge may underlie the HD patients’
poor performance, a proposal that speaks well to the observed diYculties in the
Making Inferences subtest in the present study. Once again, patients with vascular
NTS lesions showed a remarkably similar performance on subtests where an indeterminacy of meaning exists, suggesting a common subcortical basis for the
impairment in these aspects of language processing.
Conclusions
Patients with HD in the present study demonstrated deŽ cits on complex language
tasks primarily involving lexico–semantic operations, on both single-word and
sentence-level generative tasks, and on tasks which required interpretation of
ambiguous, Ž gurative and inferential meaning. These Ž ndings can be generalized,
however, only to that group of people with HD who are living in the community
and attend a local support group. The diYculties that patients with HD experienced
with tasks assessing complex language abilities were similar, both qualitatively and
quantitatively, to the language proŽ le produced by NTS subjects. Whilst the language
deŽ cits in HD subjects may be associated with speciŽ c neurocognitive impairments
in, for example, set-shifting and negation of contexts, the results raise the possibility
that a signature language proŽ le exists in both HD and NTS patients.
472
H. J. Chenery et al.
Acknowledgements
This research was supported by a grant from the Australian Research Council No.
A79600223 to H. C. and B. M. The authors thank Dr Ross Darnell for invaluable
statistical advice, Kerrin Finch, Tessa Barnett and Louise Cahill for assistance in
data collection, and the staV of the Huntington’s Association of Queensland who
assisted in recruitment. Grateful thanks are also extended to the subjects who
participated in the study.
References
Alexander, M. P., Naeser, M. A. and Palumbo, C. L., 1987, Correlations of subcortical CT lesion sites
and aphasia proŽ les. Brain, 110, 961–991.
Barr, A. and Brandt, J., 1996, Word-list generation deŽ cits in dementia. Journal of Clinical and Experimental
Neuropsychology, 6, 810–822.
Brouwers, P., Cox, C., Martin, A., Chase, T. and Fedio, P., 1984, DiVerential perceptual–spatial
impairment in Huntington’s and Alzheimner’s dementias. Archives of Neurology, 41, 1073–1076.
Butters, M. A., Goldstein, G., Allen, D. N. and Shemansky, W. J., 1998, Neuropsychological similarities
and diVerences among Huntington’s disease, multiple sclerosis and cortical dementia. Archives of
Clinical Neuropsychology, 13, 721–735.
Butters, N., Sax, D., Montgomery, K. and Tarlow, S., 1978, Comparison of the neuropsychological
deŽ cits associated with early and advanced Huntington’s disease. Archives of Neurology, 35, 585–589.
Caine, E. D., Bamford, K. A., Schiffer, R. B., Shoulson, I. and Levy, S., 1986, A controlled
neuropsychological comparison of Huntington’s disease and multiple sclerosis. Archives of Neurology,
43, 249–254.
Copland, D. A., Chenery, H. J. and Murdoch, B. E., 2000, Persistent deŽ cits in complex language
function following dominant nonthalamic subcortical lesions. Journal of Medical Speech–Language
Pathology, 8, 1–14.
D’Esposito, M. and Alexander, M. P., 1995, Subcortical aphasia: distinct proŽ les following left putaminal
hemorrhage. Neurology, 45, 38–41.
Folstein, S. E., Jensen, B., Leigh, R. J. and Folstein, M. F., 1983, The measurement of abnormal
movement: Methods developed for Huntington’s disease. Neurobehavioural Toxicology and Teratology,
5, 605–609.
Frank, E. M., McDade, H. L. and Scott, W. K., 1996, Naming in dementia secondary to Parkinson’s,
Huntington’s and Alzheimer’s diseases. Journal of Communication Disorders, 29, 183–197.
Godefroy, O., Rousseaux, M., Pruvo, J. P., Cabaret, M. and Leys, D., 1994, Neuropsychological changes
related to unilateral lenticulostriate infarcts. Journal of Neurology, Neurosurgery and Psychiatry, 57, 480–485.
Hanes, K. R., Andrewes, D. G. and Pantelis, C., 1995, Cognitive  exibility and complex integration in
Parkinson’s disease, Huntington’s disease, and schizophrenia. Journal of the International Neuropsychological
Society, 1, 545–553.
Hodges, J. R., Salmon, D. P. and Butters, N., 1991, The nature of the naming deŽ cit in Alzheimer’s
and Huntington’s disease. Brain, 114, 1547–1558.
Huisingh, R., Barrett, M. and Zachman, L., 1990, The Word Test—Revised (Moline, Illinois: Lingui
Systems).
Jaccard, J. and Wan, C. K., 1996, LISREL Approaches to Interaction EVects in Multiple Regression (Thousand
Oaks: Sage).
Kaplan, E., Goodglass, H. and Weintraub, S., 1983, The Boston Naming Test (Philadelphia: Lea & Febiger).
Kennedy, M. and Murdoch, B. E., 1993, Chronic aphasia subsequent to striato-capsular and thalamic
lesions in the left hemisphere. Brain and Language, 44, 284–295.
Kertesz, A., 1982, The Western Aphasia Battery (New York: Grune & Stratton).
Lawrence, A. D., Sahakian, B. J. and Robbins, T. W., 1998, Cognitive functions and corticostriatal
circuits: insights from Huntington’s disease. Trends in Cognitive Sciences, 2, 379–388.
Mattis, S., 1988, Dementia Rating Scale (Odessa, Florida: Psychological Assessment Resources).
Mega, M. S. and Alexander, M. P., 1994, Subcortical aphasia: the core proŽ le of capsulostriatal infarction.
Neurology, 44, 1824–1829.
Murray, L. L. and Stout, J. C., 1999, Discourse comprehension in Huntington’s and Parkinson’s diseases.
American Journal of Speech–Language Pathology, 8, 137–148.
Podoll, K., Caspary, P., Lange, H. W. and Noth, J., 1988, Language functions in Huntington’s disease.
Brain, 111, 1475–1503.
Language function in Huntington’s disease
473
Robin, D. A. and Schienberg, S., 1990, Subcortical lesions and aphasia. Journal of Speech and Hearing
Disorders, 55, 90–100.
Rosser, A. and Hodges, J. R., 1994, Initial letter and semantic category  uency in Alzheimer’s disease,
Huntington’s disease and progressive supranuclear palsy. Journal of Neurology, Neurosurgery and Psychiatry,
57, 1389–1394.
Selemon, L. D., Rajkowska, G. and Goldman-Rakic, P. S., 1998, Elevated neuronal density in prefrontal
area 46 in brains from schizophrenic patients: Application of a three-dimensional, stereologic
counting method. Journal of Comparative Neurology, 392, 402–412.
Shoulson, I. and Fahn, S., 1979, Huntington’s disease: clinical care and evaluation. Neurolog y, 29, 1–3.
Tabachnick, B. and Fidell, L., 1995, Using Multivariate Statistics (New York: Harper Collins College).
Vallar, G., Papagno, C. and Cappa, S., 1988, Latent dysphasia after left hemisphere lesions: a lexical–
semantic and verbal memory deŽ cit. Aphasiology, 2, 463–478.
Wallesch, C. W. and Fehrenbach, R. A., 1988, On the neurolinguistic nature of language abnormalities
in Huntington’s disease. Journal of Neurology, Neurosurgery and Psychiatry, 51, 367–373.
Wallesch, C. W., Kornhuber, H. H., Brunner, R. J., Kunz, T., Hollerbach, B. and Sugar, G., 1983,
Lesions of the basal ganglia, thalamus and deep white matter: diVerential eVects on language
functions. Brain and Language, 20, 286–304.
Wallesch, C. W. and Papagno, C., 1988, Subcortical aphasia. In F. C. Rose, R. Whurr and M. A. Wyke
(eds), Aphasia (London: Whurr), pp. 256–287.
White, R. W., Vasterling, J. J., Koroshetz, W. and Myers, R., 1992, Neuropsychology of Huntington’s
disease. In R. J. White (ed.), Clinical Syndromes in Adult Neuropsychology: The Practitioner’s Handbook
(New York: Elsevier).
Wiig, E. H. and Secord, W., 1989, Test of Language Competence—Expanded (New York: Psychological
Corporation).
Wiig, E. H. and Secord, W., 1992, Test of Word Knowledge (San Antonio, Texas: Psychological Corporation).
Zakzanis, K. K., 1998, The subcortical dementia of Huntington’s disease. Journal of Clinical and Experimental
Neuropsychology, 20, 565–578.
Appendix: Explanation of subtests from Battery 2
Subtest
Description
Example
Test of Language Competence—Expanded (Wiig and Secord 1989)
Ambiguous sentences
This subtest assesses the
Subjects are presented with
subject’s ability to recognize and written sentences that have two
interpret the alternative meanings alternative meanings and are
of ambiguous sentences using
asked to give, Ž rst, one meaning
both lexical and structural
of the sentence and then another
ambiguities
meanings. For, example, The
elephant was ready to lift or Jane had
a bad d ay when she broke her heel
Listening comprehension:
making inferences
The purpose of this task is to
test ability of a subject to make
permissible inferences on the
basis of existing causal
relationships or chains in short
paragraphs
The subject is read two
statements that are also shown in
print, and is required to select
two statements that best explain
what could have happened. The
subject is given four choices
from which to select
Oral expression: recreating
speech acts/recreating
sentences
The ability of a subject to
formulate propositions in
grammatically complete
sentences, incorporating
keywords related to a situation
or context is assessed in this
subtest
The subject is shown a pictured
scene (e.g. people in a park) and
is given three words that are
printed on a page (e.g. sit
painted because). The subject is
asked to provide a sentence
using all the words that could
have been used in that context
474
Figurative language
H. J. Chenery et al.
In this subtest, subjects are asked
to interpret metaphoric
expressions and then match
structurally related metaphoric
expressions by shared meaning
Test of Word Knowledge (Wiig and Secord 1992)
Word opposites
A subject’s knowledge of word
opposites is assessed in this
subtest
The subject is read a context
sentence, e.g. The situation is two
boys talking at a dog show. One of
them said ‘He is crazy about that
pet.’ What did the boy mean? After
the subject has expressed in
his/her own words what the boy
meant, they are then show four
written sentences, with the
instruction to point to the
sentence that you could use
instead of ‘He is crazy about that
pet’
The subject is given a word, e.g.
day and is asked to point to the
word that is opposite in meaning
from a choice of three words
Word deŽ nitions
This subtest assesses a subject’s
ability to provide deŽ nitions that
include category membership
and semantic features
A subject is given a word, e.g. jug
and predator and is asked to say
what kind of thing it is and then
tell some things about the word
Synonyms
A subject is assessed on their
knowledge of synonyms or
relational word knowledge
Using a written multiple-choice
format, the subject is asked to
point to a word from an array of
four words (e.g. consult embarrass
pardon pester), that means the
same as a target word (e.g. bother)
Multiple contexts
Having been given a word that
has two meanings, the subject is
asked to describe the two
contexts or meanings
The subject is given a word that
has multiple meanings, e.g. foot
and mole and is asked to provide
two meanings
The Word Test—R (Huisingh et al. 1990)
Semantic absurdities
This subtest assesses a subject’s
capacity for verbal reasoning and
the ability to recognize critical
semantic features of vocabulary
The subject is given a sentence,
e.g. The mother fed the lullaby to her
baby and is asked to say how the
sentence could be changed to
make sense