BRAIN
AND
LANGUAGE
43,
386-399 (1992)
Impairment of Temporal Organization of Speech in Basal Ganglia Diseases J. VOLKMANN,* H. HEFTER,* H. W. LANGE,? AND H.-J. FREUND* *Department
of Neurology
and the TDepartment of Psychiatry, Diisseldorf, Germany
Heinrich-Heine
University,
Absolute and relative speech timing were examined in patients suffering from Parkinson’s, Huntington’s, and Wilson’s disease. The task was to speak a standard sentence 10 times, first slowly, and then successively faster up to maximum rate. All patient groups had low maximal speech rates and showed decreased variability of speech rate. The duration of pauses between words was the same as in normals and the relative time structure of the test sentence was basically preserved. For comparison, two cases with nonfluent aphasia had even slower speech rates, large increases in pause duration, and major changes in relative speech timing. The results show the same type of alterations of the temporal organization of speech as those characteristic for rapid alternating limb movements in such patients. They support the view that the speech and skeletomotor systems share common neural control modes despite fundamental biomechanical differences. The common denominator between the speech and the skeletomotor disturbances in basal ganglia diseases may be the undamping and slowing of a fast central oscillator. o 1992 Academic Press, Inc.
INTRODUCTION The disturbance of time control of self-initiated voluntary limb movements is one of the common and well-known symptoms in all basal ganglia (BG) diseases. Rapid single and fast alternating movements are slowed and disturbed by irregularities (Hefter, Hornberg, Lange, & Freund, 1987; Hefter, Hornberg, & Freund, 1988). In normal subjects the fastest voluntary alternating hand movements are limited to the peak frequency of physiological tremor of the corresponding limb (Freund, 1983). For Parkinson’s disease (PD) Logigian, Address correspondence and reprint requests to Jens Volkmann, Department of Neurology, Heimich-Heine-University, Moorenstr. 5, 4000 Dusseldorf, Germany. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (SFB 194, A5) and a personal grant to Jens Volkmann from the Studienstiftung des Deutschen Volkes. 386 0093-934X/92 $5.OO
Copyright Q 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
SPEECH IN BASAL GANGLIA
DISEASES
387
Hefter, Reuters, and Freund (1991) have shown a concomitant slowing of both: serial alternating limb movements and tremor, which is thought to be one aspect of bradykinesia. A second problem also afflicting the temporal characteristics of movement execution arises from the “hastening phenomenon” in PD. This is the difficulty in producing auditory-paced tapping rates between 2 and 5 Hz that are different from the tremor rate. Obviously, the strong synchronization underlying pathological tremor in PD serves as an “attractor” on repetitive voluntary motor behavior. The concomitant frequency alterations of voluntary and involuntary motor phenomena in BG diseases raise the question of whether they indicate a causal relationship between their neural generators. Speech represents another alternating movement with grossly different mechanical properties (Ostry, Cooke, & Munhall, 1987). Despite this, preferred frequencies for speech (Goldmann-Eisler, 1956; Malecot, Johnston, & Kizziar, 1972; Grosjean, & Deschamps, 1973; Miller, Grosjean, & Lomanto, 1984) and handwriting or typing (Teulings, 1989; Stemberg, Monsell, Knoll, & Wright, 1978) are similar, so that these complex language-related movements all lie in the 4- to ~-HZ range. Former studies have shown that spontaneous speech rate and speech rate during reading of a text are normal or increased in PD (Canter, 1963; Ludlow, Connor, & Bassich, 1987; van Lancker & Canter, 1981; for review see Ackermann & Ziegler, 1989). This would imply that speech is excluded from the general process of bradykinesia, although “micrographia” and “hypophonia” as amplitude disturbances affect the skeletomotor system and the vocal tract. Allen (1973) has forwarded the hypothesis that two kinds of time control must underlie the generation of speech, a “global time control,” setting the time intervals during which a small number of syllables have to be produced, and a “local time control,” setting the relative timing among these basic time intervals and being necessary also for amplitude effects such as pronounciation and intonation. These two levels of time control are not totally independent and underlie an inverse relationship. We have chosen a new approach for examining speech timing in BG disorders by focusing on the ability to vary speech rate from slow to the fastest possible rates. In addition we have examined the relative timing pattern of the speech sequence, including the pauses between successive syllables and words. Using this new method we indeed show a general speech slowing in BG diseases. In order to demonstrate a differential effect of a general slowing of speech movements on the one hand and the impairment of the relative time structure of speech on the other hand, the results from patients with BG diseaseswere compared to those from patients suffering from Broca’s aphasia.
388
VOLKMANN
ET AL.
METHODS Subjects. Patients with the three major basal ganglia diseases were tested: 9 patients with Parkinson’s disease (PD), 13 patients with Huntington’s disease (HD), and 4 patients with Wilson’s disease (WD). All PD patients (three male, six female; age range: 48 to 71 years; mean age: 60.8 ? 8.49 years) suffered from an akinetic-rigid form of ideopathic PD and were receiving Ldopa treatment. In half the cases additional anticholinergic drugs were given. Five of them had on-off fluctuations. The actual impairment of the patients was assessedusing the Webster Rating Scale. The scores ranged between 3 and 2.5 (mean 15 + 7.02). Duration of disease was at least 3 years (mean 7.57 + 3.06 years). If on-off phenomena occurred, patients were tested during the off phase. The HD patients (seven male, six female; age range: 29 to 61 years; mean age: 48.1 + 9.8 years) were in different stages of the disease. Scoring of impairment during everyday life activities [modified version (Lange, Strauss, Hassel, Weller, & Tegeler, 1983) of the Shoulson-Fahn Score (Shot&on & Fahn, 1979)] was between 0 and 2 (mean 1.25 + 0.95). This score has recently been shown to correlate significantly with the degree of dementia and the degree of atrophy of the caudate head (Lange & Aulich, 1986; Starkstein, Brandt, Folstein, Strauss, Berthier, Pearlson, Wong, McDonnel, & Folstein, 1988; Bamford, Caine, Kido, Plassche, & Shoulson, 1989). Chorea score, indicating severity of chorea during clinical investigation (Lange et al., 1983) was between 1 and 2.5 (mean 1.61 + 0.48).’ Four patients presenting only mild motor dysfunctions were without therapy, the rest of the HD patients received neuroleptic drugs. All patients were able to understand the task. The WD patients (one male, three female; age range: 20 to 58 years; mean age: 40.6 ? 16.6) were taking D-penicillamine (mean dose: 1200 mg). In the three female patients presenting mild neurological symptoms, WD had been diagnosed for at least several years. The male patient was newly diagnosed and severely affected when he underwent our testing procedure. As a control group, nine subjects (seven male, two female) without any history of speech or neurological disorders were analyzed, most of them from the staff of our department. The age range was 19 to 49 years (mean age: 32.3 f 10.16 years). Testing and analytic procedure. Patients were instructed to repeat the German sentence “Fabian sass im Cafe” (Fabian sat in the coffee house)-being the modified first sentence of the novel Fabian written by Erich KLstner (1931)-10 times with different speed. In the first trial the sentence had to be said slowly. Then speaking rate had to be increased sentence by sentence, finally reaching maximal individual speech rate. The task was only explained to the patients and controls but not demonstrated to avoid any imitative effects or preference of a certain onset speech rate. Two complete runs for each subject were recorded on a high-quality tape recorder (Aiwa AD F 360). The oscillographic output of these voice signals (Figs. la and lb) was displayed off-line on a Nicolet MED 80 computer. In cursor interaction the following timing parameters were measured for each of the 20 trials of a patient: total sentence duration, pause durations, and interpauses utterance durations. The following quantities were calculated from the data: Speech rate (SR in syllables per second or Hz) as the quotient of total sentence duration and the number of syllables in our test sentence (n = 7); articulation rate (AR in syllables per second or Hz) as the quotient of total sentence duration discounting time of pauses and the number of syllables in the test sentence (n = 7); variability of speech rate (VarSR in syllables per second or Hz) as the difference between maximal and minimal speech rates of all trials of a subject; pause percentage (PP in percentage): percentage time or proportional length of all pauses in the test sentence.
’ A score of 3 is given if severe hyperkinesia is observed.
SPEECH IN BASAL GANGLIA
a
za:s
‘fa:bian
NS: slow
‘fabjan
zas im
ka’fe:
im
speech
389
DISEASES
rate
ka’fe:
Patient
(WD):
slow
speech
rate
-
Patient
(WD):
fast
speech
rate
-
163
ms
82 ms
FIG. 1. (a) Phonograms of a normal subject (NS) saying the German sentence “Fabian sass im Cafe” with the same intonation at different speech rates. Phonetic transcription is printed above the plots. (b) Phonograms of a WD patient suffering from a mild dysarthria. A slow and a fast production of the test sentence are shown. Two of the authors (J.V. and H.H.) independently evaluated the patient records for dysarthria. The following grading was used: no, mild, moderate, and severe dysarthria. The articulatory diadochokinesia was additionally tested in nine selected subjects. The method was previously described by Hefter et al. (1988) and is based on the analysis of maximimal repetition rate of the syllable “ta.”
390
VOLKMANN
ET AL.
For evaluation of relative speech timing all trials of a subject were ranked by total sentence duration. The durations of the measured syllables (“Fa,” “bian,” “sass,” “im,” “Cafe”) and all pauses were normalized to total sentence duration and plotted on a relative scale according to their occurrence in the sentence (Fig. 6). The resulting integrated sequential interval histograms showed the relative speech timing pattern of the test sentence and were compared for normal subjects and patients. Stati.rtics. A nonparametric procedure (Mann-Whitney-Wilcoxon U test) was applied for group comparison. The relationship between variables was assessed using Pearson’s rank correlation coefficient and linear regression analysis.
RESULTS Speech Timing Parameters Slowing of speech was the major finding in all patients regardless of whether they were suffering from a hypokinetic-rigid or hyperkinetic type of basal ganglia disease (Table 1, Fig. 2). Comparison with the control group yielded highly significant reduction of SR and AR (Table 1). Maximal speech rate (Max.SR) was significantly reduced in all patient groups. PD patients tended to be less affected than WD patients and HD patients (Table 1; Fig. 3). The relationship between maximal speech rate and articulatory diadochokinesia was examined by computing Pearson’s correlation coefficient. Our assumption of a strong positive correlation was confirmed (r = .88), although maximal speech rate showed higher values than maximal repetition rate of the syllable “ta” in all cases. The self-chosen minimal speech rate (Min.SR) was higher in PD and HD patients, whereas WD patients tended to have lower values of minimal speech rate than the control group, but all testing failed to be significant (Table 1). The reduction of maximal speech rate in combination with a normal or increased minimal speech rate led to a loss of Var.SR in all patients and was maximal in HD (Table 1; Fig. 3). Speech slowing in BG diseases cannot be explained by increasing pause duration since the patient groups even tended to have lower values of PP than normal subjects but most of the results were not significant (Table 1). Moreover AR was highly correlated to speech rate in the control as well as in the patient groups (NS, r = .99; HD, r = .99; PD, r = .96; WD; r = .99), whereas correlations between PP and SR were weak (NS, r= - .41; HD, r = .29; PD, r = - .03; WD, r = - .23). As can be seen in the scattergraphs in which PP was plotted against total sentence duration (Fig. 4), in normal subjects PP was only increased when total sentence duration reached values of more than 2000 msec (equivalent to a speech rate of less than 3.5 Hz). The same strategy of slowing speech by increasing PP was observed in two of the PD patients, but was missing in all WD and HD patients. Correlations of the duration of the measured speech segments and total
SPEECH IN BASAL GANGLIA
391
DISEASES
TABLE 1 SPEECH TIMING
Control group
HD patients
PD patients
WD patients
PARAMETERS OF THE PATIENT
GROWS AND THE CONTROL GROUP
Timing parameter
Mean
SR (Hz) AR (Hz) PP (%) Max.SR (Hz) Min.SR (Hz) Var.SR (Hz)
5.51 5.96 8.15 9.07 2.75 6.31
SR W AR (Hz) PP (%) MaxSR (Hz) Min.SR (Hz) Var.SR (Hz)
4.27 4.61 7.15 5.42 3.34 2.09
(0.94) (1.05) (2.72) (1.W (0.85) (1.28)
ns. .ooo n.s. .ooo
SR (Hz) AR (Hz) PP (%) Max.SR (Hz) Min.SR (Hz) Var.SR (Hz)
4.69 5.08 8.04 6.52 3.08 3.44
(1.39) (1.47) (5.62) (1.17) (1.28) (1.31)
.ooo .ooo n.s. ,003 n.s. .003
.ooo .ooo
SR (Hz) AR (Hz) PP (%) Max.SR (Hz) Min.SR (Hz) VarSR (Hz)
4.09 4.44 8.14 5.96 2.20 3.76
(1.30) (1.36) (2.71) (0.46) (0.67) (0.45)
.ooo .ooo ,007 .006 n.s. ,009
n.s. n.s. .003 n.s. n.s. n.s.
ww
PW)
PWD)
PPD)
(2.03)
G-9)
(4.55) (1.W (0.77) (1.97)
.ooo .mo
n.s. n.s. n.s. n.s. .ooo .ooo n.s. n.s. n.s. ns.
Note. SR, speech rate; AR, articulation rate; PP, pause percentage; Min.SR, minimal speech rate; Max.SR, maximal speech rate; Var.SR, variability of speech rate; n.s., not significant; p(NS), p(two-tailed) for group comparison with control group; p(HD), p(twotailed) for group comparison with HD patients; p(PD), p(two-tailed) for group comparison with PD patients.
sentence duration indicated that speech rate was mainly controlled by phoneme duration (Fig. 5). Relationship between Speech Timing Measures and Clinical Aspects For the HD patients the scoring of impairment during everyday life activities [modified version (Lange et al., 1983) of the Shoulson-Fahn Score (Shoulson & Fahn, 1979)J was correlated with mean speech rate. We found a significant negative correlation (r = - .64; p < .05). A significant correlation was also found between increasing disability scores and decrease of maximal speech rate (r = - .73; p < .Ol). Similar correlations for PD patients between the Webster score and speech timing
392
VOLKMANN
ET AL.
I
probability,/““, /’ / 0.3
/ ?
,, “‘ /
/
/
0,1 0.05 0 .25
1.25 2.25 3.25 4.25 5.25 6.25 7.25 6.25 9.2510.25
speech rate (syllables/set) FIG. 2. The histogram shows the distribution of speech rates produced in all trials by the control group (NS) and the patient groups (WD, Wilson’s disease; PD, Parkinson’s disease; HD, Huntington’s disease). The preferred speech rates can be identified from the relative maxima in the distributions.
parameters showed the same tendency but did not reach levels of significance. Evaluation of all patient records for dysarthria yielded the following
2! 12345676910 number of sentence
repetition
FIG. 3. Increase of speech rate over 10 sentence repetitions is shown for the control group (NS) and the patient groups (HD, WD, PD) with linear regression lines computed separately for each group. In the patients maximal speech rate and variability of speech rate are affected to different extents.
SPEECH IN BASAL GANGLIA
393
DISEASES
-=j$$...Y~.~~; :: ,P~-j .+..i;
AND
LANGUAGE
43,
386-399 (1992)
Impairment of Temporal Organization of Speech in Basal Ganglia Diseases J. VOLKMANN,* H. HEFTER,* H. W. LANGE,? AND H.-J. FREUND* *Department
of Neurology
and the TDepartment of Psychiatry, Diisseldorf, Germany
Heinrich-Heine
University,
Absolute and relative speech timing were examined in patients suffering from Parkinson’s, Huntington’s, and Wilson’s disease. The task was to speak a standard sentence 10 times, first slowly, and then successively faster up to maximum rate. All patient groups had low maximal speech rates and showed decreased variability of speech rate. The duration of pauses between words was the same as in normals and the relative time structure of the test sentence was basically preserved. For comparison, two cases with nonfluent aphasia had even slower speech rates, large increases in pause duration, and major changes in relative speech timing. The results show the same type of alterations of the temporal organization of speech as those characteristic for rapid alternating limb movements in such patients. They support the view that the speech and skeletomotor systems share common neural control modes despite fundamental biomechanical differences. The common denominator between the speech and the skeletomotor disturbances in basal ganglia diseases may be the undamping and slowing of a fast central oscillator. o 1992 Academic Press, Inc.
INTRODUCTION The disturbance of time control of self-initiated voluntary limb movements is one of the common and well-known symptoms in all basal ganglia (BG) diseases. Rapid single and fast alternating movements are slowed and disturbed by irregularities (Hefter, Hornberg, Lange, & Freund, 1987; Hefter, Hornberg, & Freund, 1988). In normal subjects the fastest voluntary alternating hand movements are limited to the peak frequency of physiological tremor of the corresponding limb (Freund, 1983). For Parkinson’s disease (PD) Logigian, Address correspondence and reprint requests to Jens Volkmann, Department of Neurology, Heimich-Heine-University, Moorenstr. 5, 4000 Dusseldorf, Germany. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (SFB 194, A5) and a personal grant to Jens Volkmann from the Studienstiftung des Deutschen Volkes. 386 0093-934X/92 $5.OO
Copyright Q 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
SPEECH IN BASAL GANGLIA
DISEASES
387
Hefter, Reuters, and Freund (1991) have shown a concomitant slowing of both: serial alternating limb movements and tremor, which is thought to be one aspect of bradykinesia. A second problem also afflicting the temporal characteristics of movement execution arises from the “hastening phenomenon” in PD. This is the difficulty in producing auditory-paced tapping rates between 2 and 5 Hz that are different from the tremor rate. Obviously, the strong synchronization underlying pathological tremor in PD serves as an “attractor” on repetitive voluntary motor behavior. The concomitant frequency alterations of voluntary and involuntary motor phenomena in BG diseases raise the question of whether they indicate a causal relationship between their neural generators. Speech represents another alternating movement with grossly different mechanical properties (Ostry, Cooke, & Munhall, 1987). Despite this, preferred frequencies for speech (Goldmann-Eisler, 1956; Malecot, Johnston, & Kizziar, 1972; Grosjean, & Deschamps, 1973; Miller, Grosjean, & Lomanto, 1984) and handwriting or typing (Teulings, 1989; Stemberg, Monsell, Knoll, & Wright, 1978) are similar, so that these complex language-related movements all lie in the 4- to ~-HZ range. Former studies have shown that spontaneous speech rate and speech rate during reading of a text are normal or increased in PD (Canter, 1963; Ludlow, Connor, & Bassich, 1987; van Lancker & Canter, 1981; for review see Ackermann & Ziegler, 1989). This would imply that speech is excluded from the general process of bradykinesia, although “micrographia” and “hypophonia” as amplitude disturbances affect the skeletomotor system and the vocal tract. Allen (1973) has forwarded the hypothesis that two kinds of time control must underlie the generation of speech, a “global time control,” setting the time intervals during which a small number of syllables have to be produced, and a “local time control,” setting the relative timing among these basic time intervals and being necessary also for amplitude effects such as pronounciation and intonation. These two levels of time control are not totally independent and underlie an inverse relationship. We have chosen a new approach for examining speech timing in BG disorders by focusing on the ability to vary speech rate from slow to the fastest possible rates. In addition we have examined the relative timing pattern of the speech sequence, including the pauses between successive syllables and words. Using this new method we indeed show a general speech slowing in BG diseases. In order to demonstrate a differential effect of a general slowing of speech movements on the one hand and the impairment of the relative time structure of speech on the other hand, the results from patients with BG diseaseswere compared to those from patients suffering from Broca’s aphasia.
388
VOLKMANN
ET AL.
METHODS Subjects. Patients with the three major basal ganglia diseases were tested: 9 patients with Parkinson’s disease (PD), 13 patients with Huntington’s disease (HD), and 4 patients with Wilson’s disease (WD). All PD patients (three male, six female; age range: 48 to 71 years; mean age: 60.8 ? 8.49 years) suffered from an akinetic-rigid form of ideopathic PD and were receiving Ldopa treatment. In half the cases additional anticholinergic drugs were given. Five of them had on-off fluctuations. The actual impairment of the patients was assessedusing the Webster Rating Scale. The scores ranged between 3 and 2.5 (mean 15 + 7.02). Duration of disease was at least 3 years (mean 7.57 + 3.06 years). If on-off phenomena occurred, patients were tested during the off phase. The HD patients (seven male, six female; age range: 29 to 61 years; mean age: 48.1 + 9.8 years) were in different stages of the disease. Scoring of impairment during everyday life activities [modified version (Lange, Strauss, Hassel, Weller, & Tegeler, 1983) of the Shoulson-Fahn Score (Shot&on & Fahn, 1979)] was between 0 and 2 (mean 1.25 + 0.95). This score has recently been shown to correlate significantly with the degree of dementia and the degree of atrophy of the caudate head (Lange & Aulich, 1986; Starkstein, Brandt, Folstein, Strauss, Berthier, Pearlson, Wong, McDonnel, & Folstein, 1988; Bamford, Caine, Kido, Plassche, & Shoulson, 1989). Chorea score, indicating severity of chorea during clinical investigation (Lange et al., 1983) was between 1 and 2.5 (mean 1.61 + 0.48).’ Four patients presenting only mild motor dysfunctions were without therapy, the rest of the HD patients received neuroleptic drugs. All patients were able to understand the task. The WD patients (one male, three female; age range: 20 to 58 years; mean age: 40.6 ? 16.6) were taking D-penicillamine (mean dose: 1200 mg). In the three female patients presenting mild neurological symptoms, WD had been diagnosed for at least several years. The male patient was newly diagnosed and severely affected when he underwent our testing procedure. As a control group, nine subjects (seven male, two female) without any history of speech or neurological disorders were analyzed, most of them from the staff of our department. The age range was 19 to 49 years (mean age: 32.3 f 10.16 years). Testing and analytic procedure. Patients were instructed to repeat the German sentence “Fabian sass im Cafe” (Fabian sat in the coffee house)-being the modified first sentence of the novel Fabian written by Erich KLstner (1931)-10 times with different speed. In the first trial the sentence had to be said slowly. Then speaking rate had to be increased sentence by sentence, finally reaching maximal individual speech rate. The task was only explained to the patients and controls but not demonstrated to avoid any imitative effects or preference of a certain onset speech rate. Two complete runs for each subject were recorded on a high-quality tape recorder (Aiwa AD F 360). The oscillographic output of these voice signals (Figs. la and lb) was displayed off-line on a Nicolet MED 80 computer. In cursor interaction the following timing parameters were measured for each of the 20 trials of a patient: total sentence duration, pause durations, and interpauses utterance durations. The following quantities were calculated from the data: Speech rate (SR in syllables per second or Hz) as the quotient of total sentence duration and the number of syllables in our test sentence (n = 7); articulation rate (AR in syllables per second or Hz) as the quotient of total sentence duration discounting time of pauses and the number of syllables in the test sentence (n = 7); variability of speech rate (VarSR in syllables per second or Hz) as the difference between maximal and minimal speech rates of all trials of a subject; pause percentage (PP in percentage): percentage time or proportional length of all pauses in the test sentence.
’ A score of 3 is given if severe hyperkinesia is observed.
SPEECH IN BASAL GANGLIA
a
za:s
‘fa:bian
NS: slow
‘fabjan
zas im
ka’fe:
im
speech
389
DISEASES
rate
ka’fe:
Patient
(WD):
slow
speech
rate
-
Patient
(WD):
fast
speech
rate
-
163
ms
82 ms
FIG. 1. (a) Phonograms of a normal subject (NS) saying the German sentence “Fabian sass im Cafe” with the same intonation at different speech rates. Phonetic transcription is printed above the plots. (b) Phonograms of a WD patient suffering from a mild dysarthria. A slow and a fast production of the test sentence are shown. Two of the authors (J.V. and H.H.) independently evaluated the patient records for dysarthria. The following grading was used: no, mild, moderate, and severe dysarthria. The articulatory diadochokinesia was additionally tested in nine selected subjects. The method was previously described by Hefter et al. (1988) and is based on the analysis of maximimal repetition rate of the syllable “ta.”
390
VOLKMANN
ET AL.
For evaluation of relative speech timing all trials of a subject were ranked by total sentence duration. The durations of the measured syllables (“Fa,” “bian,” “sass,” “im,” “Cafe”) and all pauses were normalized to total sentence duration and plotted on a relative scale according to their occurrence in the sentence (Fig. 6). The resulting integrated sequential interval histograms showed the relative speech timing pattern of the test sentence and were compared for normal subjects and patients. Stati.rtics. A nonparametric procedure (Mann-Whitney-Wilcoxon U test) was applied for group comparison. The relationship between variables was assessed using Pearson’s rank correlation coefficient and linear regression analysis.
RESULTS Speech Timing Parameters Slowing of speech was the major finding in all patients regardless of whether they were suffering from a hypokinetic-rigid or hyperkinetic type of basal ganglia disease (Table 1, Fig. 2). Comparison with the control group yielded highly significant reduction of SR and AR (Table 1). Maximal speech rate (Max.SR) was significantly reduced in all patient groups. PD patients tended to be less affected than WD patients and HD patients (Table 1; Fig. 3). The relationship between maximal speech rate and articulatory diadochokinesia was examined by computing Pearson’s correlation coefficient. Our assumption of a strong positive correlation was confirmed (r = .88), although maximal speech rate showed higher values than maximal repetition rate of the syllable “ta” in all cases. The self-chosen minimal speech rate (Min.SR) was higher in PD and HD patients, whereas WD patients tended to have lower values of minimal speech rate than the control group, but all testing failed to be significant (Table 1). The reduction of maximal speech rate in combination with a normal or increased minimal speech rate led to a loss of Var.SR in all patients and was maximal in HD (Table 1; Fig. 3). Speech slowing in BG diseases cannot be explained by increasing pause duration since the patient groups even tended to have lower values of PP than normal subjects but most of the results were not significant (Table 1). Moreover AR was highly correlated to speech rate in the control as well as in the patient groups (NS, r = .99; HD, r = .99; PD, r = .96; WD; r = .99), whereas correlations between PP and SR were weak (NS, r= - .41; HD, r = .29; PD, r = - .03; WD, r = - .23). As can be seen in the scattergraphs in which PP was plotted against total sentence duration (Fig. 4), in normal subjects PP was only increased when total sentence duration reached values of more than 2000 msec (equivalent to a speech rate of less than 3.5 Hz). The same strategy of slowing speech by increasing PP was observed in two of the PD patients, but was missing in all WD and HD patients. Correlations of the duration of the measured speech segments and total
SPEECH IN BASAL GANGLIA
391
DISEASES
TABLE 1 SPEECH TIMING
Control group
HD patients
PD patients
WD patients
PARAMETERS OF THE PATIENT
GROWS AND THE CONTROL GROUP
Timing parameter
Mean
SR (Hz) AR (Hz) PP (%) Max.SR (Hz) Min.SR (Hz) Var.SR (Hz)
5.51 5.96 8.15 9.07 2.75 6.31
SR W AR (Hz) PP (%) MaxSR (Hz) Min.SR (Hz) Var.SR (Hz)
4.27 4.61 7.15 5.42 3.34 2.09
(0.94) (1.05) (2.72) (1.W (0.85) (1.28)
ns. .ooo n.s. .ooo
SR (Hz) AR (Hz) PP (%) Max.SR (Hz) Min.SR (Hz) Var.SR (Hz)
4.69 5.08 8.04 6.52 3.08 3.44
(1.39) (1.47) (5.62) (1.17) (1.28) (1.31)
.ooo .ooo n.s. ,003 n.s. .003
.ooo .ooo
SR (Hz) AR (Hz) PP (%) Max.SR (Hz) Min.SR (Hz) VarSR (Hz)
4.09 4.44 8.14 5.96 2.20 3.76
(1.30) (1.36) (2.71) (0.46) (0.67) (0.45)
.ooo .ooo ,007 .006 n.s. ,009
n.s. n.s. .003 n.s. n.s. n.s.
ww
PW)
PWD)
PPD)
(2.03)
G-9)
(4.55) (1.W (0.77) (1.97)
.ooo .mo
n.s. n.s. n.s. n.s. .ooo .ooo n.s. n.s. n.s. ns.
Note. SR, speech rate; AR, articulation rate; PP, pause percentage; Min.SR, minimal speech rate; Max.SR, maximal speech rate; Var.SR, variability of speech rate; n.s., not significant; p(NS), p(two-tailed) for group comparison with control group; p(HD), p(twotailed) for group comparison with HD patients; p(PD), p(two-tailed) for group comparison with PD patients.
sentence duration indicated that speech rate was mainly controlled by phoneme duration (Fig. 5). Relationship between Speech Timing Measures and Clinical Aspects For the HD patients the scoring of impairment during everyday life activities [modified version (Lange et al., 1983) of the Shoulson-Fahn Score (Shoulson & Fahn, 1979)J was correlated with mean speech rate. We found a significant negative correlation (r = - .64; p < .05). A significant correlation was also found between increasing disability scores and decrease of maximal speech rate (r = - .73; p < .Ol). Similar correlations for PD patients between the Webster score and speech timing
392
VOLKMANN
ET AL.
I
probability,/““, /’ / 0.3
/ ?
,, “‘ /
/
/
0,1 0.05 0 .25
1.25 2.25 3.25 4.25 5.25 6.25 7.25 6.25 9.2510.25
speech rate (syllables/set) FIG. 2. The histogram shows the distribution of speech rates produced in all trials by the control group (NS) and the patient groups (WD, Wilson’s disease; PD, Parkinson’s disease; HD, Huntington’s disease). The preferred speech rates can be identified from the relative maxima in the distributions.
parameters showed the same tendency but did not reach levels of significance. Evaluation of all patient records for dysarthria yielded the following
2! 12345676910 number of sentence
repetition
FIG. 3. Increase of speech rate over 10 sentence repetitions is shown for the control group (NS) and the patient groups (HD, WD, PD) with linear regression lines computed separately for each group. In the patients maximal speech rate and variability of speech rate are affected to different extents.
SPEECH IN BASAL GANGLIA
393
DISEASES
-=j$$...Y~.~~; :: ,P~-j .+..i;
