- Email: [email protected]
Is the Paced Auditory Serial Addition Test (Pasat) a Valid Means of Assessing Executive Function in Parkinson's Disease?
RESEARCH REPORT IS THE PACED AUDITORY SERIAL ADDITION TEST (PASAT) A VALID MEANS OF ASSESSING EXECUTIVE FUNCTION IN PARKINSON’S DISEASE? Kathy Dujardin1, Caroline Denève1, Mélanie Ronval1, Pierre Krystkowiak1, Christine Humez2, Alain Destée1 and Luc Defebvre1 (1Neurology and Movement Disorders Unit, Faculty of Medicine and Lille University Hospital, Lille, France; 2Psychology Department, Charles De Gaulle University, Lille, France)
ABSTRACT Parkinson’s disease (PD) is usually associated with a dysexecutive syndrome. However, many executive function tasks require visuo-spatial abilities which themselves are known to be impaired in PD. The use of a non-visual procedure may thus represent a means of avoiding this type of methodological difficulty. The aim of the present study was to determine whether the paced auditory serial addition test (PASAT) might constitute a useful procedure for assessing executive functions in PD. Twenty-seven non-demented PD patients early in the course of the disease participated in the study, together with 15 healthy control (HC) subjects. All participants performed the PASAT and a set of clinical tasks assessing information processing speed, working memory and executive functions. Compared with HCs, the PD patients were significantly impaired in their performance of the PASAT. Significant impairment (compared with controls) was also evidenced by only one of the clinical tasks – the symbol coding task, which assesses information processing speed. Our results demonstrate the high sensitivity of the PASAT to cognitive impairment. However, correlation analyses showed that the main factor explaining the PD patients’ PASAT impairment was cognitive slowing. Key words: basal ganglia, cognitive slowing
INTRODUCTION Parkinson’s disease (PD) is a neurodegenerative disorder whose main clinical manifestations include rest tremor, rigidity, akinesia, and postural instability. The disease is usually associated with cognitive deficits: overall cognitive efficiency is generally either preserved or only mildly impaired, and the deficits mainly concern tasks involving executive abilities (Pillon et al., 1996). Even though a certain proportion of patients will always meet the criteria for dementia (usually late in the course of the disease), the main cognitive impairments in non-demented PD patients concern tasks requiring planning and goal-directed behaviour (Owen et al., 1992), set-shifting (Owen et al., 1993), inhibition of a preferred response (Brown and Marsden, 1988; Dujardin et al., 1999) and action coordination (Dalrymple-Alford et al., 1994; Malapani et al., 1994). This subcorticofrontal syndrome is observed from the first stages of the disease onwards and its severity usually increases with disease progression (Owen et al., 1995). Most of the tasks used to assess executive function in PD are based on visual material and thus require visuo-spatial abilities. However, it has been shown that visuospatial processing is disturbed in PD: although the matter is still subject to debate, the impairment apparently concerns tasks requiring high attentional control or self-initiated response Cortex, (2007) 43, 601-606
strategies (Brown and Marsden, 1990; Flowers and Robertson, 1995; Crevits and De Ridder, 1997). Since the use of a non-visual task could represent a way of avoiding this difficulty, the aim of the present study was to determine whether the paced auditory serial addition test (PASAT) might constitute a useful procedure for assessing executive function in PD. The PASAT was initially devised by Gronwall (1977) in order to provide an estimate of the subject’s rate of information processing. However, it also requires the subject to i) actively maintain and update information in working memory, ii) efficiently inhibit encoding of his/her own response while attending to the next stimulus in the series and iii) perform a task according to an externally paced rhythm. Sustained attention, working memory and executive functions thus appear to be involved, in addition to information processing speed (Diehr et al., 1998). In clinical practice, the PASAT is not frequently used with PD patients. It is, however, among the tests recommended by the core assessment program for surgical interventional therapies in Parkinson’s disease (CAPSIT-PD; Defer et al. 1999) and its use has been proposed by some research teams for the evaluation of attention, working memory and processing speed in PD (Saint-Cyr and Trepanier, 2000). However, a more accurate appraisal of the PASAT’s validity in this pathology seems necessary. We therefore compared the PASAT
602
Kathy Dujardin and Others TABLE I
Demographical characteristics of the participant groups (mean (SD) and [range])
Number Male/Female Age Education duration (years) Mattis DRS (/144) MADRS (/60) UPDRS-III (/108) off drug Disease duration (years)
Untreated PD patients
Treated PD patients
Healthy controls
14 8/6 62 (8) [50-80] 12 (3) [8-19] 139.57 (3.52) [134-144] 8.57 (4.67) [2-14] 18.67 (8.31) [3-29] 1.57 (1.24) [.5-4]
13 8/5 63 (11) [40-80] 11 (2.5) [8-16] 138.69 (4.05) [132-143] 4.69 (3.22) [0-9] 20.20 (6.76) [11-27] 4.38 (3.12) [2-12]
15 6/9 61 (9) [46-77] 11 (3) [9-20] 140.40 (2.90) [133-144] 1.27 (1.28) [0-4]
F (2, 38)
.146 .943 .813 17.567*
*p < .01.
performance of PD patients early in the course of the disease with that of healthy control (HC) subjects, using two presentation rates. We predicted that performance in the PASAT would correlate with other measures of executive function.
Although none of the participants met the diagnostic criteria for depression, both groups of PD patients had a mean MADRS score significantly higher than the HC subjects (p < .001). The three groups did not differ in terms of the other demographic variables (see Table I).
METHODS
Tasks
Participants Twenty-seven non-demented patients with probable PD (11 females, 16 males) participated in the study. PD was defined according to the criteria of the United Kingdom Parkinson’s Disease Brain Bank (Gibb and Lees, 1988). Fourteen patients were early in the course of PD and had not yet received any medication, whereas thirteen were receiving antiparkinsonian medication and presented a stable clinical picture. Motor disability was evaluated using the motor score of the unified idiopathic Parkinson’s disease rating scale (UPDRS-III), varying from 0 (optimal score) to 108 (worst score) (Fahn et al., 1987). Fifteen HC subjects (9 females and 6 males, chosen to match the patient group as closely as possible with respect to age and education) also participated in the study. They had no history of neurological or psychiatric illness, and their family history was negative for PD or parkinsonian symptoms. None of the participants were receiving drugs which could potentially interfere with cognitive function. Participants were required to show a Mattis dementia rating scale (DRS) score higher than the lowest quintile of the reference population (Schmidt et al., 1994) and a Montgomery and Asberg depression rating scale (MADRS) score lower than 15 (the recommended cut-off for PD patients (Leentjens et al., 2000). Low ability in performing elementary additions was also an exclusion criterion. The characteristics of the patients and control subjects are shown in Table I.
The Paced Auditory Serial Addition Test (PASAT) The test was designed according to the procedure described by Gronwall (1977). Series of 61 numbers from 1 to 9 were randomly delivered at presentation rates of one number every 2.8 or 1.6 sec. Each series was preceded by a practice list of 10 numbers delivered at the same presentation rate. Each number from 1 to 9 was first recorded as an audio file whose duration was .5 sec. Thereafter, two different series of 61 numbers were pseudo-randomly constituted. The presentation was controlled by software especially developed for the study. Participants were instructed to add each number to the one immediately preceding it: the second had to be added to the first, the third to the second, and so on. Performance was assessed in terms of the percentage of correct additions. The participants completed a set of clinical tasks (assessing processing speed, working memory and executive function) during the same session as the PASAT task. Symbol Coding In this task, nine nonsense symbols were randomly presented in rows. A key was printed above these rows, showing each nonsense symbol paired with a number. The participant’s task was to say (as quickly as possible and without error) the number corresponding to each symbol. The task lasted 90 sec and performance was assessed in terms of the number of symbols correctly coded.
Early cognitive slowing in PD
603
TABLE II
Percentage of correct additions in the PASAT (mean (SD) and [range])
2.8 sec interstimulus interval 1.6 sec interstimulus interval
Untreated PD patients
Treated PD patients
Healthy controls
78.81 (17.49) [45-98.33] 57.17 (17.21) [31.67-82]
75.51 (20.02) [50-100] 60.77 (20.28) [21.67-96.67]
93.44 (8.74) [75-100] 83.11 (13.05) [60-100]
Digit Span The WAIS-R digit span subtest yielded the participant’s forward and backward digit spans. Digit Ordering Task In this task, digit sequences were orally presented to the participants. The size of the sequences was adjusted to each participant’s forward span. They were instructed to repeat the sequence after ordering the digits in ascending order. The task comprised ten trials. Performance was assessed in terms of the percentage of correct responses.
MADRS score as a covariate) to test for the effect of group and presentation rate on the percentage of correct additions in the PASAT. The significance level was set at .05. The same analyses were carried out to test for the effect of group on the dependent measures of the clinical tests. A conservative criterion of p < .01 was chosen for these analyses, in order to lessen the likelihood of a type 1 error caused by multiple comparisons. Spearman correlation coefficients were calculated in order to investigate the relationship between performance in the clinical tests and the PASAT. Again, in view of potential multiple correlations, a conservative criterion of p < .01 was chosen.
Stroop Word-Colour Task RESULTS A 50-item version of the test was applied. The procedure (which has been fully described elsewhere (Dujardin et al., 2004)) comprised two trials – a baseline condition and an interference condition. Performance was assessed in terms of the time in seconds needed to complete each phase, together with an interference cost index. Letter and Number Sequencing Task This task (consisting of an oral version of the Trail Making Test) has been described fully elsewhere (Dujardin et al., 2000). Performance was evaluated in terms of an alternation cost index. Crossed Tapping Test This test was designed by Godefroy et al. (1992). Participants were given a stick and were instructed to listen to a sound recording. When they heard a single, brief sound, they had to tap twice on the table with the stick; when they heard two, consecutive, brief sounds, they had to tap once. Ten practice trials were run before starting the actual task, which comprised 40 trials. Performance was assessed in terms of the number of errors.
PASAT Mean (SD) percentages of correct responses at each presentation rate are presented in Table II. The ANCOVA revealed a significant main effect of presentation rate [F (1, 38) = 29.72, p < .001] and group [F (2, 38) = 8.34, p = .001]. There was a trend toward a significant group x presentation rate interaction [F (2, 38) = 3.03, p = .06]. MADRS depression score was not a significant covariate [F (1, 38) = .63, p = .43]. Performance was always better at a low presentation rate. Post-hoc analyses revealed that treated PD patients performed significantly worse than HCs (p = .02) with the 2.8 sec interstimulus interval. For untreated PD patients, there was a trend toward performing worse than HCs (p = .058). There was no difference between the two patient groups (p = .87). With the 1.6 sec interval, both treated (p = .005) and untreated (p = .001) PD patients performed significantly worse than HCs, and there was no difference between the two patient groups (p = .86). Clinical Tasks Performance (mean (SD)) is presented in Table III.
Data Analysis Since the groups differed in terms of the mean MADRS depression score, we used univariate (ANCOVA) and, when appropriate, multivariate (MANCOVA) analyses of covariance (with the
The MANCOVA yielded a trend toward a significant effect of group [F (18, 60) = 2.16, p = .014] on performance in the clinical tasks. MADRS depression score was not a significant covariate [F (9, 30) = 1.09, p = .39]. Further analyses revealed a significant main effect of group on one parameter
604
Kathy Dujardin and Others TABLE III
Performance scores (mean (SD) and [range]) for the clinical tasks Untreated PD patients
Treated PD patients
Healthy controls
F (3, 41)
44.07 (6.34) [32-53] 5.36 (.84) [4-7] 3.78 (.97) [2-5]
43.08 (8.90) [29-56] 5.92 (1.11) [4-8] 3.00 (.91) [2-5]
55.00 (8.31) [39-65] 6.00 (.84) [5-7] 3.73 (1.33) [2-6]
6.55*
90.89 (12.66) [52-100]
83.92 (13.30) [44-96.66]
87.32 (9.83) [67.14-100]
2.17
35.54 (7.74) [26-55] 26.14 (16.74) [0-65]
32.54 (2.90) [25-36] 20.08 (6.89) [10-32]
30.4 (5.60) [25-43] 18.4 (5.90) [10-32]
1.92
20.27 (9.52) [6-39]
21.61 (10.33) [9-49]
15.47 (4.66) [7-25]
2.40
1.36 (2.84) [0-10]
2.53 (2.78) [0-8]
.67 (1.34) [0-5]
2.14
Symbol coding Symbols correctly coded Forward digit span Backward digit span Digit ordering Correct responses (%) Stroop word colour test Time to complete phase 1 (sec) Interference cost index Letter number sequencing task Alternation cost index Crossed tapping test Errors
1.33 1.48
1.33
*p < .01.
only: the number of symbols correctly coded in the symbol coding task [F (3, 41) = 6.55, p = .001]. There was no significant group effect on the other parameters (see Table III). Post-hoc analyses revealed that both treated (p = .001) and untreated (p = .002) PD patients performed significantly worse than HCs in the symbol coding task. There was no significant difference between the two patient groups (p = .95).
PASAT performance by either patients or HCs did not correlate with the Mattis DRS score, the backward digit span, the percentage of correct responses in the digit ordering task or the interference index in the Stroop word-colour test. In HCs, PASAT performance merely tended to correlate with the time to complete the baseline condition of the Stroop word-colour test and the processing speed in the symbol coding task. In the patient group, PASAT performance significantly correlated not only with these two latter parameters but also with the alternating cost index in the letter/number sequencing task and the number of errors in the crossed tapping task. Furthermore, there was a trend toward a significant correlation with the forward digit span. This suggests thus that PASAT performance in PD patients is related to both general
Correlation Analyses Since our analysis had not revealed any significant differences between the two PD patient groups, correlations were calculated for the patient group as a whole. The results are presented in Table IV.
TABLE IV
Correlation between performance in the PASAT and in the clinical tasks (Spearman Rho coefficients) Healthy controls
Mattis DRS Symbol coding Symbols correctly coded Forward digit span Backward digit span Digit ordering Correct responses (%) Stroop word colour test Time to complete phase 1 (sec) Interference cost index Letter number sequencing task Alternation cost index Crossed tapping test Errors *p < .05, **p < .01.
PD patients
PASAT 2.8
PASAT 1.6
PASAT 2.8
.417
.425
.608* – .164 .232
.565* .455 .156
.635** .316 .098
.694** .489* .372
.505
.380
.281
.160
– .593* – .373
– .637* .027
– .821** – .092
– .625** – .145
.016
– .123
– .476*
– .578**
– .079
.118
– .523**
– .349
– .062
PASAT 1.6 .133
Early cognitive slowing in PD TABLE V
Partial correlations between PD patients’ performances in the PASAT and in the clinical tasks (with processing speed in the symbol coding task held constant) PD patients PASAT 2.8 Mattis DRS Symbol coding Forward digit span Backward digit span Digit ordering Correct responses (%) Stroop word colour test Time to complete phase 1 (sec) Interference cost index Letter number sequencing task Alternation cost index Crossed tapping test Errors
PASAT 1.6
– .193
.003
.239 – .031
.238 .311
.211
.176
– .464* – .015
– .269 – .088
– .203
– .259
– .254
– .078
* p < .05, **p < .01.
information processing speed (as in HCs) and setshifting abilities. However, this latter assumption appears difficult to justify, since our PD patients did not show deficits in any of the tasks assessing executive function in general and set-shifting in particular (see Table III). Since the only difference between both groups concerned the symbol coding task, it is possible that the correlation between PASAT performance and the parameters assessing set-shifting abilities is due to cognitive slowing in PD patients. Partial correlations (with processing speed in the symbol coding task held constant) were thus performed (see Table V). This analysis showed that PASAT performance was correlated with the time to complete the baseline condition of the Stroop word-colour test but not with any of the other parameters. DISCUSSION The present results demonstrate the PASAT’s high sensitivity to cognitive impairment. Indeed, with patients early in the course of PD and lacking overt cognitive decline, an impairment in PASAT performance was evident – even at a low presentation rate. However, despite its high sensitivity, our results do not validate the PASAT as a measure of executive function in PD. In fact, the pattern of correlations between performance in the PASAT and in the other clinical tasks reveals that PASAT performance could mainly be explained by information processing speed: after controlling for the influence of this factor, no correlation with the other measures of executive function was observed. Moreover, besides the PASAT, only symbol coding (a widely accepted measure of information processing speed) was significantly impaired in PD patients. This finding agrees with the assumption of Gronwall (1977) that PASAT performance generally assesses rate of
605
information processing, and thus suggests that the PD patients’ decline in PASAT performance is merely an index of their cognitive slowing. Consequently, despite its high sensitivity to deficit, the PASAT does not seem to provide more information than the symbol coding task. Since the latter is self-paced, easy to administer and (in the form used here) does not require motor participation (thus ruling out contamination by motor slowing), it seems more suitable than the PASAT for assessing cognitive slowing early in the course of PD. Alternatively, since depression and cognitive slowing are related in PD (Rogers et al., 1987), the general slowing in information processing we observed in PD patients might follow mild symptoms of depression. However, such an interpretation seems unlikely for our results both because we carefully excluded depressive patients and because, despite a higher MADRS mean score in the PD patients, this factor was never a significant covariate in our statistical analyses. In conclusion, the PASAT does not enable the assessment of specific executive dysfunction in PD. The PD patients’ PASAT impairment appears to be a consequence of their cognitive slowing. This slowing occurs very early in the course of the disease and is insensitive to dopaminergic treatment, since no differences between treated and untreated patients were observed. REFERENCES BROWN RG and MARSDEN CD. An investigation of the phenomenon of “set” in Parkinson’s disease. Movement Disorders, 3: 152-161, 1988. BROWN RG and MARSDEN CD. Cognitive function in Parkinson’s disease: From description to theory. Trends in Neurosciences, 13: 21-29, 1990. CREVITS L and DE RIDDER K. Disturbed striatoprefrontal mediated visual behavior in moderate to severe parkinsonian patients. Journal of Neurology, Neurosurgery and Psychiatry, 63: 296299, 1997. DALRYMPLE-ALFORD JC, KALDERS AS, JONES RD and WATSON RW. A central executive deficit in patients with Parkinson’s disease. Journal of Neurology, Neurosurgery and Psychiatry, 57: 360-367, 1994. DEFER GL, WIDNER H, MARIÉ RM, REMY P and LEVIVIER M. Core assessment program for surgical interventional therapies in Parkinson’s disease (CAPSIT-PD). Movement Disorders, 14: 572-584, 1999. DIEHR MC, HEATON RK, MILLER W and GRANT I. The Paced Auditory Serial Addition Task (PASAT): Norms for age, education, and ethnicity. Assessment, 5: 375-387, 1998 [Erratum in Assessment, 6: 101, 1999]. DUJARDIN K, BLAIRY S, DEFEBVRE L, DUHEM S, NOEL Y, HESS U and DESTÉE A. Deficits in decoding emotional facial expressions in Parkinson’s disease. Neuropsychologia, 42: 239-250, 2004. DUJARDIN K, DEGREEF JF, ROGELET P, DEFEBVRE L and DESTÉE A. Impairment of the supervisory attentional system in early untreated patients with Parkinson’s disease. Journal of Neurology, 246: 783-788, 1999. DUJARDIN K, KRYSTKOWIAK P, DEFEBVRE L, BLOND S and DESTÉE A. A case of severe dysexecutive syndrome consecutive to chronic bilateral pallidal stimulation. Neuropsychologia, 38: 1305-1315, 2000. FAHN S, ELTON RL and MEMBERS OF THE UPDRS DEVELOPMENT COMMITTEE. Unified idiopathic Parkinson’s disease rating scale. In Fahn S, Marsden CD, Calne D and Goldstein M (Eds), Recent Developments in Parkinson’s Disease (vol. 2). Florham Park: Mac Millan Healthcare Information, 1987.
606
Kathy Dujardin and Others
FLOWERS KA and ROBERTSON C. Perceptual abnormalities in Parkinson’s disease: Top-down or bottom-up processes? Perception, 24: 1201-1221, 1995. GIBB WR and LEES AJ. The prevalence of the Lewy body to the pathogenesis of idiopathic Parkinson’s disease. Journal of Neurology, Neurosurgery and Psychiatry, 51: 745-752, 1988. GODEFROY O, ROUSSEAUX M, LEYS D, DESTÉE A, SCHELTENS P and PRUVO JP. Frontal lobe dysfunction in unilateral lenticulostriate infarcts. Prominent role of cortical lesions. Archives of Neurology, 49: 1285-1289, 1992. GRONWALL DM. Paced auditory serial addition task: A measure of recovery from concussion. Perceptual and Motor Skills, 44: 367-373, 1977. LEENTJENS AF, VERHEY FR, LOUSBERG R, SPITSBERGEN H and WILMINK FW. The validity of the Hamilton and MontgomeryAsberg depression rating scales as screening and diagnostic tools for depression in Parkinson’s disease. International Journal of Geriatric Psychiatry, 15: 644-649, 2000. MALAPANI C, PILLON B, DUBOIS B and AGID Y. Impaired simultaneous cognitive task performance in Parkinson’s disease: A dopamine-related dysfunction. Neurology, 44: 319326, 1994. OWEN AM, JAMES M, LEIGH PN, SUMMERS BA, MARSDEN CD, QUINN NP, LANGE KW and ROBBINS TW. Fronto-striatal cognitive deficits at different stages of Parkinson’s disease. Brain, 115: 1727-1751, 1992.
OWEN AM, ROBERTS AC, HODGES JR, SUMMERS BA, POLKEY CE and ROBBINS TW. Contrasting mechanisms of impaired attentional set-shifting in patients with frontal lobe damage or Parkinson’s disease. Brain, 116: 1159-1175, 1993. OWEN AM, SAHAKIAN BJ, HODGES JR, SUMMERS BA, POLKEY CE and ROBBINS TW. Dopamine-dependent frontostriatal planning deficits in early Parkinson’s disease. Neuropsychology, 9: 126-140, 1995. PILLON B, DUBOIS B and AGID Y. Testing cognition may contribute to the diagnosis of movement disorders. Neurology, 46: 324334, 1996. ROGERS D, LEES AJ, SMITH E, TRIMBLE M and STERN GM. Bradyphrenia in Parkinson’s disease and psychomotor retardation in depressive illness. An experimental study. Brain, 110: 761-776, 1987. SAINT-CYR JA and TREPANIER LL. Neuropsychologic assessment of patients for movement disorder surgery. Movement Disorders, 15: 771-783, 2000. SCHMIDT R, FRIEDL W, FAZEKAS F, REINHART B, GRIESHOPER P, KOCH M, SCHAKER M, POLMIN K and LECHNER H. The Mattis dementia rating scale: Normative data from 1001 healthy volunteers. Neurology, 44: 964-966, 1994. Kathy Dujardin, Neurologie et Pathologie du Mouvement, Neurologie A, Hôpital Salengro, Centre Hospitalier Universitaire, F-59037 Lille Cedex, France. e-mail: [email protected]
(Received 29 October 2004; reviewed 4 February 2005; revised 30 May 2005; accepted 17 August 2005; Action Editor Paolo Nichelli)
ABSTRACT Parkinson’s disease (PD) is usually associated with a dysexecutive syndrome. However, many executive function tasks require visuo-spatial abilities which themselves are known to be impaired in PD. The use of a non-visual procedure may thus represent a means of avoiding this type of methodological difficulty. The aim of the present study was to determine whether the paced auditory serial addition test (PASAT) might constitute a useful procedure for assessing executive functions in PD. Twenty-seven non-demented PD patients early in the course of the disease participated in the study, together with 15 healthy control (HC) subjects. All participants performed the PASAT and a set of clinical tasks assessing information processing speed, working memory and executive functions. Compared with HCs, the PD patients were significantly impaired in their performance of the PASAT. Significant impairment (compared with controls) was also evidenced by only one of the clinical tasks – the symbol coding task, which assesses information processing speed. Our results demonstrate the high sensitivity of the PASAT to cognitive impairment. However, correlation analyses showed that the main factor explaining the PD patients’ PASAT impairment was cognitive slowing. Key words: basal ganglia, cognitive slowing
INTRODUCTION Parkinson’s disease (PD) is a neurodegenerative disorder whose main clinical manifestations include rest tremor, rigidity, akinesia, and postural instability. The disease is usually associated with cognitive deficits: overall cognitive efficiency is generally either preserved or only mildly impaired, and the deficits mainly concern tasks involving executive abilities (Pillon et al., 1996). Even though a certain proportion of patients will always meet the criteria for dementia (usually late in the course of the disease), the main cognitive impairments in non-demented PD patients concern tasks requiring planning and goal-directed behaviour (Owen et al., 1992), set-shifting (Owen et al., 1993), inhibition of a preferred response (Brown and Marsden, 1988; Dujardin et al., 1999) and action coordination (Dalrymple-Alford et al., 1994; Malapani et al., 1994). This subcorticofrontal syndrome is observed from the first stages of the disease onwards and its severity usually increases with disease progression (Owen et al., 1995). Most of the tasks used to assess executive function in PD are based on visual material and thus require visuo-spatial abilities. However, it has been shown that visuospatial processing is disturbed in PD: although the matter is still subject to debate, the impairment apparently concerns tasks requiring high attentional control or self-initiated response Cortex, (2007) 43, 601-606
strategies (Brown and Marsden, 1990; Flowers and Robertson, 1995; Crevits and De Ridder, 1997). Since the use of a non-visual task could represent a way of avoiding this difficulty, the aim of the present study was to determine whether the paced auditory serial addition test (PASAT) might constitute a useful procedure for assessing executive function in PD. The PASAT was initially devised by Gronwall (1977) in order to provide an estimate of the subject’s rate of information processing. However, it also requires the subject to i) actively maintain and update information in working memory, ii) efficiently inhibit encoding of his/her own response while attending to the next stimulus in the series and iii) perform a task according to an externally paced rhythm. Sustained attention, working memory and executive functions thus appear to be involved, in addition to information processing speed (Diehr et al., 1998). In clinical practice, the PASAT is not frequently used with PD patients. It is, however, among the tests recommended by the core assessment program for surgical interventional therapies in Parkinson’s disease (CAPSIT-PD; Defer et al. 1999) and its use has been proposed by some research teams for the evaluation of attention, working memory and processing speed in PD (Saint-Cyr and Trepanier, 2000). However, a more accurate appraisal of the PASAT’s validity in this pathology seems necessary. We therefore compared the PASAT
602
Kathy Dujardin and Others TABLE I
Demographical characteristics of the participant groups (mean (SD) and [range])
Number Male/Female Age Education duration (years) Mattis DRS (/144) MADRS (/60) UPDRS-III (/108) off drug Disease duration (years)
Untreated PD patients
Treated PD patients
Healthy controls
14 8/6 62 (8) [50-80] 12 (3) [8-19] 139.57 (3.52) [134-144] 8.57 (4.67) [2-14] 18.67 (8.31) [3-29] 1.57 (1.24) [.5-4]
13 8/5 63 (11) [40-80] 11 (2.5) [8-16] 138.69 (4.05) [132-143] 4.69 (3.22) [0-9] 20.20 (6.76) [11-27] 4.38 (3.12) [2-12]
15 6/9 61 (9) [46-77] 11 (3) [9-20] 140.40 (2.90) [133-144] 1.27 (1.28) [0-4]
F (2, 38)
.146 .943 .813 17.567*
*p < .01.
performance of PD patients early in the course of the disease with that of healthy control (HC) subjects, using two presentation rates. We predicted that performance in the PASAT would correlate with other measures of executive function.
Although none of the participants met the diagnostic criteria for depression, both groups of PD patients had a mean MADRS score significantly higher than the HC subjects (p < .001). The three groups did not differ in terms of the other demographic variables (see Table I).
METHODS
Tasks
Participants Twenty-seven non-demented patients with probable PD (11 females, 16 males) participated in the study. PD was defined according to the criteria of the United Kingdom Parkinson’s Disease Brain Bank (Gibb and Lees, 1988). Fourteen patients were early in the course of PD and had not yet received any medication, whereas thirteen were receiving antiparkinsonian medication and presented a stable clinical picture. Motor disability was evaluated using the motor score of the unified idiopathic Parkinson’s disease rating scale (UPDRS-III), varying from 0 (optimal score) to 108 (worst score) (Fahn et al., 1987). Fifteen HC subjects (9 females and 6 males, chosen to match the patient group as closely as possible with respect to age and education) also participated in the study. They had no history of neurological or psychiatric illness, and their family history was negative for PD or parkinsonian symptoms. None of the participants were receiving drugs which could potentially interfere with cognitive function. Participants were required to show a Mattis dementia rating scale (DRS) score higher than the lowest quintile of the reference population (Schmidt et al., 1994) and a Montgomery and Asberg depression rating scale (MADRS) score lower than 15 (the recommended cut-off for PD patients (Leentjens et al., 2000). Low ability in performing elementary additions was also an exclusion criterion. The characteristics of the patients and control subjects are shown in Table I.
The Paced Auditory Serial Addition Test (PASAT) The test was designed according to the procedure described by Gronwall (1977). Series of 61 numbers from 1 to 9 were randomly delivered at presentation rates of one number every 2.8 or 1.6 sec. Each series was preceded by a practice list of 10 numbers delivered at the same presentation rate. Each number from 1 to 9 was first recorded as an audio file whose duration was .5 sec. Thereafter, two different series of 61 numbers were pseudo-randomly constituted. The presentation was controlled by software especially developed for the study. Participants were instructed to add each number to the one immediately preceding it: the second had to be added to the first, the third to the second, and so on. Performance was assessed in terms of the percentage of correct additions. The participants completed a set of clinical tasks (assessing processing speed, working memory and executive function) during the same session as the PASAT task. Symbol Coding In this task, nine nonsense symbols were randomly presented in rows. A key was printed above these rows, showing each nonsense symbol paired with a number. The participant’s task was to say (as quickly as possible and without error) the number corresponding to each symbol. The task lasted 90 sec and performance was assessed in terms of the number of symbols correctly coded.
Early cognitive slowing in PD
603
TABLE II
Percentage of correct additions in the PASAT (mean (SD) and [range])
2.8 sec interstimulus interval 1.6 sec interstimulus interval
Untreated PD patients
Treated PD patients
Healthy controls
78.81 (17.49) [45-98.33] 57.17 (17.21) [31.67-82]
75.51 (20.02) [50-100] 60.77 (20.28) [21.67-96.67]
93.44 (8.74) [75-100] 83.11 (13.05) [60-100]
Digit Span The WAIS-R digit span subtest yielded the participant’s forward and backward digit spans. Digit Ordering Task In this task, digit sequences were orally presented to the participants. The size of the sequences was adjusted to each participant’s forward span. They were instructed to repeat the sequence after ordering the digits in ascending order. The task comprised ten trials. Performance was assessed in terms of the percentage of correct responses.
MADRS score as a covariate) to test for the effect of group and presentation rate on the percentage of correct additions in the PASAT. The significance level was set at .05. The same analyses were carried out to test for the effect of group on the dependent measures of the clinical tests. A conservative criterion of p < .01 was chosen for these analyses, in order to lessen the likelihood of a type 1 error caused by multiple comparisons. Spearman correlation coefficients were calculated in order to investigate the relationship between performance in the clinical tests and the PASAT. Again, in view of potential multiple correlations, a conservative criterion of p < .01 was chosen.
Stroop Word-Colour Task RESULTS A 50-item version of the test was applied. The procedure (which has been fully described elsewhere (Dujardin et al., 2004)) comprised two trials – a baseline condition and an interference condition. Performance was assessed in terms of the time in seconds needed to complete each phase, together with an interference cost index. Letter and Number Sequencing Task This task (consisting of an oral version of the Trail Making Test) has been described fully elsewhere (Dujardin et al., 2000). Performance was evaluated in terms of an alternation cost index. Crossed Tapping Test This test was designed by Godefroy et al. (1992). Participants were given a stick and were instructed to listen to a sound recording. When they heard a single, brief sound, they had to tap twice on the table with the stick; when they heard two, consecutive, brief sounds, they had to tap once. Ten practice trials were run before starting the actual task, which comprised 40 trials. Performance was assessed in terms of the number of errors.
PASAT Mean (SD) percentages of correct responses at each presentation rate are presented in Table II. The ANCOVA revealed a significant main effect of presentation rate [F (1, 38) = 29.72, p < .001] and group [F (2, 38) = 8.34, p = .001]. There was a trend toward a significant group x presentation rate interaction [F (2, 38) = 3.03, p = .06]. MADRS depression score was not a significant covariate [F (1, 38) = .63, p = .43]. Performance was always better at a low presentation rate. Post-hoc analyses revealed that treated PD patients performed significantly worse than HCs (p = .02) with the 2.8 sec interstimulus interval. For untreated PD patients, there was a trend toward performing worse than HCs (p = .058). There was no difference between the two patient groups (p = .87). With the 1.6 sec interval, both treated (p = .005) and untreated (p = .001) PD patients performed significantly worse than HCs, and there was no difference between the two patient groups (p = .86). Clinical Tasks Performance (mean (SD)) is presented in Table III.
Data Analysis Since the groups differed in terms of the mean MADRS depression score, we used univariate (ANCOVA) and, when appropriate, multivariate (MANCOVA) analyses of covariance (with the
The MANCOVA yielded a trend toward a significant effect of group [F (18, 60) = 2.16, p = .014] on performance in the clinical tasks. MADRS depression score was not a significant covariate [F (9, 30) = 1.09, p = .39]. Further analyses revealed a significant main effect of group on one parameter
604
Kathy Dujardin and Others TABLE III
Performance scores (mean (SD) and [range]) for the clinical tasks Untreated PD patients
Treated PD patients
Healthy controls
F (3, 41)
44.07 (6.34) [32-53] 5.36 (.84) [4-7] 3.78 (.97) [2-5]
43.08 (8.90) [29-56] 5.92 (1.11) [4-8] 3.00 (.91) [2-5]
55.00 (8.31) [39-65] 6.00 (.84) [5-7] 3.73 (1.33) [2-6]
6.55*
90.89 (12.66) [52-100]
83.92 (13.30) [44-96.66]
87.32 (9.83) [67.14-100]
2.17
35.54 (7.74) [26-55] 26.14 (16.74) [0-65]
32.54 (2.90) [25-36] 20.08 (6.89) [10-32]
30.4 (5.60) [25-43] 18.4 (5.90) [10-32]
1.92
20.27 (9.52) [6-39]
21.61 (10.33) [9-49]
15.47 (4.66) [7-25]
2.40
1.36 (2.84) [0-10]
2.53 (2.78) [0-8]
.67 (1.34) [0-5]
2.14
Symbol coding Symbols correctly coded Forward digit span Backward digit span Digit ordering Correct responses (%) Stroop word colour test Time to complete phase 1 (sec) Interference cost index Letter number sequencing task Alternation cost index Crossed tapping test Errors
1.33 1.48
1.33
*p < .01.
only: the number of symbols correctly coded in the symbol coding task [F (3, 41) = 6.55, p = .001]. There was no significant group effect on the other parameters (see Table III). Post-hoc analyses revealed that both treated (p = .001) and untreated (p = .002) PD patients performed significantly worse than HCs in the symbol coding task. There was no significant difference between the two patient groups (p = .95).
PASAT performance by either patients or HCs did not correlate with the Mattis DRS score, the backward digit span, the percentage of correct responses in the digit ordering task or the interference index in the Stroop word-colour test. In HCs, PASAT performance merely tended to correlate with the time to complete the baseline condition of the Stroop word-colour test and the processing speed in the symbol coding task. In the patient group, PASAT performance significantly correlated not only with these two latter parameters but also with the alternating cost index in the letter/number sequencing task and the number of errors in the crossed tapping task. Furthermore, there was a trend toward a significant correlation with the forward digit span. This suggests thus that PASAT performance in PD patients is related to both general
Correlation Analyses Since our analysis had not revealed any significant differences between the two PD patient groups, correlations were calculated for the patient group as a whole. The results are presented in Table IV.
TABLE IV
Correlation between performance in the PASAT and in the clinical tasks (Spearman Rho coefficients) Healthy controls
Mattis DRS Symbol coding Symbols correctly coded Forward digit span Backward digit span Digit ordering Correct responses (%) Stroop word colour test Time to complete phase 1 (sec) Interference cost index Letter number sequencing task Alternation cost index Crossed tapping test Errors *p < .05, **p < .01.
PD patients
PASAT 2.8
PASAT 1.6
PASAT 2.8
.417
.425
.608* – .164 .232
.565* .455 .156
.635** .316 .098
.694** .489* .372
.505
.380
.281
.160
– .593* – .373
– .637* .027
– .821** – .092
– .625** – .145
.016
– .123
– .476*
– .578**
– .079
.118
– .523**
– .349
– .062
PASAT 1.6 .133
Early cognitive slowing in PD TABLE V
Partial correlations between PD patients’ performances in the PASAT and in the clinical tasks (with processing speed in the symbol coding task held constant) PD patients PASAT 2.8 Mattis DRS Symbol coding Forward digit span Backward digit span Digit ordering Correct responses (%) Stroop word colour test Time to complete phase 1 (sec) Interference cost index Letter number sequencing task Alternation cost index Crossed tapping test Errors
PASAT 1.6
– .193
.003
.239 – .031
.238 .311
.211
.176
– .464* – .015
– .269 – .088
– .203
– .259
– .254
– .078
* p < .05, **p < .01.
information processing speed (as in HCs) and setshifting abilities. However, this latter assumption appears difficult to justify, since our PD patients did not show deficits in any of the tasks assessing executive function in general and set-shifting in particular (see Table III). Since the only difference between both groups concerned the symbol coding task, it is possible that the correlation between PASAT performance and the parameters assessing set-shifting abilities is due to cognitive slowing in PD patients. Partial correlations (with processing speed in the symbol coding task held constant) were thus performed (see Table V). This analysis showed that PASAT performance was correlated with the time to complete the baseline condition of the Stroop word-colour test but not with any of the other parameters. DISCUSSION The present results demonstrate the PASAT’s high sensitivity to cognitive impairment. Indeed, with patients early in the course of PD and lacking overt cognitive decline, an impairment in PASAT performance was evident – even at a low presentation rate. However, despite its high sensitivity, our results do not validate the PASAT as a measure of executive function in PD. In fact, the pattern of correlations between performance in the PASAT and in the other clinical tasks reveals that PASAT performance could mainly be explained by information processing speed: after controlling for the influence of this factor, no correlation with the other measures of executive function was observed. Moreover, besides the PASAT, only symbol coding (a widely accepted measure of information processing speed) was significantly impaired in PD patients. This finding agrees with the assumption of Gronwall (1977) that PASAT performance generally assesses rate of
605
information processing, and thus suggests that the PD patients’ decline in PASAT performance is merely an index of their cognitive slowing. Consequently, despite its high sensitivity to deficit, the PASAT does not seem to provide more information than the symbol coding task. Since the latter is self-paced, easy to administer and (in the form used here) does not require motor participation (thus ruling out contamination by motor slowing), it seems more suitable than the PASAT for assessing cognitive slowing early in the course of PD. Alternatively, since depression and cognitive slowing are related in PD (Rogers et al., 1987), the general slowing in information processing we observed in PD patients might follow mild symptoms of depression. However, such an interpretation seems unlikely for our results both because we carefully excluded depressive patients and because, despite a higher MADRS mean score in the PD patients, this factor was never a significant covariate in our statistical analyses. In conclusion, the PASAT does not enable the assessment of specific executive dysfunction in PD. The PD patients’ PASAT impairment appears to be a consequence of their cognitive slowing. This slowing occurs very early in the course of the disease and is insensitive to dopaminergic treatment, since no differences between treated and untreated patients were observed. REFERENCES BROWN RG and MARSDEN CD. An investigation of the phenomenon of “set” in Parkinson’s disease. Movement Disorders, 3: 152-161, 1988. BROWN RG and MARSDEN CD. Cognitive function in Parkinson’s disease: From description to theory. Trends in Neurosciences, 13: 21-29, 1990. CREVITS L and DE RIDDER K. Disturbed striatoprefrontal mediated visual behavior in moderate to severe parkinsonian patients. Journal of Neurology, Neurosurgery and Psychiatry, 63: 296299, 1997. DALRYMPLE-ALFORD JC, KALDERS AS, JONES RD and WATSON RW. A central executive deficit in patients with Parkinson’s disease. Journal of Neurology, Neurosurgery and Psychiatry, 57: 360-367, 1994. DEFER GL, WIDNER H, MARIÉ RM, REMY P and LEVIVIER M. Core assessment program for surgical interventional therapies in Parkinson’s disease (CAPSIT-PD). Movement Disorders, 14: 572-584, 1999. DIEHR MC, HEATON RK, MILLER W and GRANT I. The Paced Auditory Serial Addition Task (PASAT): Norms for age, education, and ethnicity. Assessment, 5: 375-387, 1998 [Erratum in Assessment, 6: 101, 1999]. DUJARDIN K, BLAIRY S, DEFEBVRE L, DUHEM S, NOEL Y, HESS U and DESTÉE A. Deficits in decoding emotional facial expressions in Parkinson’s disease. Neuropsychologia, 42: 239-250, 2004. DUJARDIN K, DEGREEF JF, ROGELET P, DEFEBVRE L and DESTÉE A. Impairment of the supervisory attentional system in early untreated patients with Parkinson’s disease. Journal of Neurology, 246: 783-788, 1999. DUJARDIN K, KRYSTKOWIAK P, DEFEBVRE L, BLOND S and DESTÉE A. A case of severe dysexecutive syndrome consecutive to chronic bilateral pallidal stimulation. Neuropsychologia, 38: 1305-1315, 2000. FAHN S, ELTON RL and MEMBERS OF THE UPDRS DEVELOPMENT COMMITTEE. Unified idiopathic Parkinson’s disease rating scale. In Fahn S, Marsden CD, Calne D and Goldstein M (Eds), Recent Developments in Parkinson’s Disease (vol. 2). Florham Park: Mac Millan Healthcare Information, 1987.
606
Kathy Dujardin and Others
FLOWERS KA and ROBERTSON C. Perceptual abnormalities in Parkinson’s disease: Top-down or bottom-up processes? Perception, 24: 1201-1221, 1995. GIBB WR and LEES AJ. The prevalence of the Lewy body to the pathogenesis of idiopathic Parkinson’s disease. Journal of Neurology, Neurosurgery and Psychiatry, 51: 745-752, 1988. GODEFROY O, ROUSSEAUX M, LEYS D, DESTÉE A, SCHELTENS P and PRUVO JP. Frontal lobe dysfunction in unilateral lenticulostriate infarcts. Prominent role of cortical lesions. Archives of Neurology, 49: 1285-1289, 1992. GRONWALL DM. Paced auditory serial addition task: A measure of recovery from concussion. Perceptual and Motor Skills, 44: 367-373, 1977. LEENTJENS AF, VERHEY FR, LOUSBERG R, SPITSBERGEN H and WILMINK FW. The validity of the Hamilton and MontgomeryAsberg depression rating scales as screening and diagnostic tools for depression in Parkinson’s disease. International Journal of Geriatric Psychiatry, 15: 644-649, 2000. MALAPANI C, PILLON B, DUBOIS B and AGID Y. Impaired simultaneous cognitive task performance in Parkinson’s disease: A dopamine-related dysfunction. Neurology, 44: 319326, 1994. OWEN AM, JAMES M, LEIGH PN, SUMMERS BA, MARSDEN CD, QUINN NP, LANGE KW and ROBBINS TW. Fronto-striatal cognitive deficits at different stages of Parkinson’s disease. Brain, 115: 1727-1751, 1992.
OWEN AM, ROBERTS AC, HODGES JR, SUMMERS BA, POLKEY CE and ROBBINS TW. Contrasting mechanisms of impaired attentional set-shifting in patients with frontal lobe damage or Parkinson’s disease. Brain, 116: 1159-1175, 1993. OWEN AM, SAHAKIAN BJ, HODGES JR, SUMMERS BA, POLKEY CE and ROBBINS TW. Dopamine-dependent frontostriatal planning deficits in early Parkinson’s disease. Neuropsychology, 9: 126-140, 1995. PILLON B, DUBOIS B and AGID Y. Testing cognition may contribute to the diagnosis of movement disorders. Neurology, 46: 324334, 1996. ROGERS D, LEES AJ, SMITH E, TRIMBLE M and STERN GM. Bradyphrenia in Parkinson’s disease and psychomotor retardation in depressive illness. An experimental study. Brain, 110: 761-776, 1987. SAINT-CYR JA and TREPANIER LL. Neuropsychologic assessment of patients for movement disorder surgery. Movement Disorders, 15: 771-783, 2000. SCHMIDT R, FRIEDL W, FAZEKAS F, REINHART B, GRIESHOPER P, KOCH M, SCHAKER M, POLMIN K and LECHNER H. The Mattis dementia rating scale: Normative data from 1001 healthy volunteers. Neurology, 44: 964-966, 1994. Kathy Dujardin, Neurologie et Pathologie du Mouvement, Neurologie A, Hôpital Salengro, Centre Hospitalier Universitaire, F-59037 Lille Cedex, France. e-mail: [email protected]
(Received 29 October 2004; reviewed 4 February 2005; revised 30 May 2005; accepted 17 August 2005; Action Editor Paolo Nichelli)