Effects of cognitive training in Parkinson's disease: A randomized controlled trial

Parkinsonism and Related Disorders 20 (2014) 1196e1202

Contents lists available at ScienceDirect

Parkinsonism and Related Disorders journal homepage: www.elsevier.com/locate/parkreldis

Effects of cognitive training in Parkinson's disease: A randomized controlled trial Annette Petrelli a, b, Stephanie Kaesberg a, b, c, Michael T. Barbe b, c, Lars Timmermann b, Gereon R. Fink b, c, Josef Kessler b, Elke Kalbe a, b, * a

Institute of Gerontology, Psychological Gerontology & Center for Neuropsychological Diagnostics and Intervention (CeNDI), University of Vechta, Driverstr. 22, 49377 Vechta, Germany Department of Neurology, University Hospital Cologne, Kerpenerstr.62, 50937, Cologne, Germany c Cognitive Neuroscience Section, Institute of Neuroscience and Medicine (INM-3), Research Center Jülich, Wilhelm-Johnen-Straße, 52428, Jülich, Germany b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 20 May 2014 Received in revised form 17 August 2014 Accepted 27 August 2014

Background: In Parkinson's Disease (PD), cognitive dysfunctions which can reduce patients' quality of life occur frequently. Data on non-pharmacological intervention effects on cognitive functions in patients with PD are rare. The aim of this study was to examine the effects of different cognitive group trainings (structured vs. unstructured) on cognition, depression, and quality of life in non-demented PD patients. Methods: In this randomized controlled trial, 65 non-demented patients with PD according to UK Brain Bank criteria (Hoehn & Yahr I-III) were allocated to one of two cognitive multi-component treatments (“NEUROvitalis”, a structured training, or the unstructured training “Mentally fit” with randomly assembled cognitive tasks, each including 12 group-sessions
a 90 min over 6 weeks) or a waiting list control group (CG). A neuropsychological test battery was performed before and after the training. Results: Compared to the CG, patients from the “NEUROvitalis” group improved in short-term memory (word list learning “Memo”: p < .01) and working memory (digit span reverse from “DemTect”: p < .05), whereas depression scores where reduced in the “Mentally fit” group (Beck Depression Inventory-II: p < .05). The “NEUROvitalis” group improved significantly more in working memory than the “Mentally fit” group (DemTect: p < .05). Discussion: Cognitive and affective functions can be improved by cognitive trainings in PD patients. Specific effects (e.g. on memory and working memory versus depression) seem to be dependent on the type of training. Further research is needed to define long-term effects and the efficacy in PD patients with different extent of cognitive and neuropsychiatric symptoms. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Parkinson's disease Cognitive training Intervention Depression

1. Introduction 1.1. Background Cognitive dysfunctions are frequent symptoms of Parkinson's disease (PD) [1e4] and can limit the patients' quality of life and their ability to organize their lives independently [5]. Since PD represents a high risk for the development of Mild Cognitive Impairment (MCI) and dementia [6] and since there is no prevention strategy for

* Corresponding author. Institute of Gerontology, Psychological Gerontology & Center for Neuropsychological Diagnostics and Intervention (CeNDI), University of Vechta, Driverstr. 22, 49377 Vechta, Germany. E-mail addresses: [email protected], [email protected] (E. Kalbe). http://dx.doi.org/10.1016/j.parkreldis.2014.08.023 1353-8020/© 2014 Elsevier Ltd. All rights reserved.

cognitive decline and no approved pharmacological approach to treat PD-MCI yet, non-pharmacological approaches have attracted increasing interest. Remarkably, there is evidence for the effectiveness of cognitive training programs in healthy subjects and (non-PD) MCI patients [4,7]. A recent review demonstrates that cognitive training (CT) and exercise training are the most frequently used non-pharmacological approaches to enhance cognitive functions in non-demented PD patients [8]. Another systematic review on the effects of exercise on cognition in PD patients shows that various types of exercise can improve cognitive function [9]. Although these studies indicate positive effects, data appeared limited. Only three CT studies were randomized controlled trials (RCTs) [10e12], and studies showed heterogeneous designs (treatment, patient characteristics, outcome measures etc.).

A. Petrelli et al. / Parkinsonism and Related Disorders 20 (2014) 1196e1202

1.2. Objectives The aim of this study was to conduct an RCT to investigate a) whether non-demented PD patients who receive a six-week cognitive group training can improve their cognitive functions (especially attention, memory, and executive functions, which are often impaired in PD patients [3]), depression scores and quality of life ratings compared to PD patients without training, and b) whether there is a difference in the effects depending on the kind of training (structured training with sessions targeting specific cognitive functions and psychoeducational elements pronouncing metacognitive knowledge versus an unstructured brain training program - similar to “brain jogging”, as it is frequently offered in “memory groups” and many books and in which cognitive tasks are randomly put together.). We hypothesized that PD patients benefit from CT (hypothesis I) and that benefits are more pronounced when taking part in a structured rather than an unstructured training (hypothesis II). 2. Methods

1197

Both treatment programs were equal with regard to the duration and the fre
90 min, 6 weeks). However, they differed quency of training sessions (12 sessions a in structure and partly also in content. Until now, there is little PD specific training available. Thus, our aim was to compare a structured training program (NV) which targets domains that are frequently impaired in PD (attention, memory, executive functions) with an unstructured, not domain-specific “brain jogging” program as it is frequently offered. The structured program NV includes individual tasks, group tasks and group games each focusing on specific cognitive functions (attention, memory, executive functions). In NV each session focuses on one specific cognitive domain or topic and starts with a corresponding psychoeducational part (e.g. cognitive functions, training possibilities, compensation strategies; see also Supplementary material). In MF, domains were not addressed in focused sessions. Individual and group tasks which trained attention, memory and less specific functions (language in general, creative thinking) were combined randomly over the course of the entire program. Instead of psychoeducation, sessions contained group conversations about topics proposed by the trainer or by the patients themselves (e.g. dealing with the disease). The exercises of MF were composed from a representative choice of tasks of eight frequently used German brain trainings in addition to one group game (see Supplementary material). To guarantee that all patients were involved in the training, patients were trained in small groups with high levels of interaction. Group tasks and games always involved all participants, and individual tasks were supervised by the trainer. Supplementary material with further information on the structure and contents of the trainings is available online.

2.1. Trial design In this RCT with two intervention groups (the structured “NEUROvitalis” [NV] [13] and the unstructured “Mentally fit” [MF]) and one waiting group (CG) PD patients were recruited from the outpatient clinic at the Movement Disorders Unit of the Department of Neurology, University Hospital of Cologne, Germany (n ¼ 40), or from regional PD support groups (n ¼ 30). After recruitment, all patients were randomly allocated with a computer program (http://www. randomizer.org); support groups: NV/MF/CG: n ¼ 7/7/16; University Hospital: NV/MF/CG: n ¼ 18/17/5). Ten training groups were built with comparable group sizes (5 NV-groups [University Hospital: n ¼ 4 þ 5 þ 8; support groups: n ¼ 3 þ 4], 5 MF-groups [University Hospital: n ¼ 3 þ 6 þ 8; support groups: n ¼ 3 þ 4]. Due to the informed consent, only patients in the treatment groups were blinded for the kind of treatment. The examinators were blinded for the condition, trainers were not. 2.2. Study setting The study took place between May 2010 and June 2012. It was conducted in accordance with the Helsinki Declaration of 1964 (2008 revision), was approved of by the local Ethics Committee of the University of Cologne (11e061) and is registered in the German Clinical Trials Register (#DRKS00004978). Written informed consent was obtained from each patient.

2.6. Neuropsychological test battery and outcome measures An elaborate neuropsychological test battery was conducted with each patient (Table 1, for references please see Supplementary material). Attention, memory (verbal short- and long-term, visual long-term), and executive functions (working memory [meaning a multi-component system that holds and manipulates information in short-term memory] and verbal fluency) which are frequently impaired in non-demented PD patients [3] and which are domains trained in both training programs were defined as primary outcome measures. Visuo-construction, depression and quality of life were defined as secondary outcomes. Patients were examined within ten days before (pre-test) and ten days after the training (post-test). In the CG, pre- and posttests were administered in a parallel time interval without training. When available, parallel test versions were applied (for DemTect [a cognitive screening tool to detect MCI and dementia with five subtests assessing verbal short- and long-term memory, working memory, semantic verbal fluency, and number transcoding], Memo, and complex figure test). 2.7. Statistical analysis

The Hoehn and Yahr (H&Y) staging of PD [15] was defined at baseline. The motor score of the Unified Parkinson's Disease Rating Scale [16] (UPDRS part III) was videotaped and rated by an MDS-certified physician (MTB) blinded for the treatment condition. The scales were rated during the “on”-phase; pre- and posttest examinations were usually conducted at the same time of the day after the neuropsychological examination.

A power analysis was performed with G*Power (http://www.psycho.uniduesseldorf.de/abteilungen/aap/gpower3/download-and-register/index_html) to determine the needed sample size for medium to large effects (partial eta-squared 2 [parts ] ¼ .06). Alpha Level was defined .05, power .80. The calculated sample size when comparing two groups was a minimum of 17 patients for each group. To compare the patients' characteristics at baseline, two-tailed t-tests for independent samples (age, education in years, MMSE, DemTect, BDI-II) were used. Additionally, patients who dropped out were compared to the group of all analyzed patients regarding these characteristics. According to our two hypotheses and as suggested by the Cochrane collaboration for studies which compare alternative interventions, we treated the study statistically as three separate trials and performed “head-tohead comparisons”: Trial 1 “NV versus CG”, Trial 2 “MF versus CG” (both hypothesis I, CT can enhance cognition), and Trial 3 “NV versus MF” (hypothesis II, structured training is more effective than unstructured training). Repeated measurement analyses of variance were performed for each trial. The repeated measurement factor in the ANOVAs was pre- versus posttest. The covariates age, education and duration of disease were included. An adjustment of p-values (.05) using Bonferroni correction was applied. Partial eta-squared was calculated as a measure of effect sizes to compare the magnitude of treatment effects (0.01 ¼ small, 0.06 ¼ medium, 0.14 ¼ large effects).

2.5. Interventions

2.8. Patient characteristics

Our aim was to conduct interventions which are easily administrable by clinicians and caring employees working with patients and therefore can be made available to a large number of patients. A group setting was selected, because (i) social activity is known to have a positive impact on cognition and the risk of dementia [17], (ii) non-demented PD patients e as compared to other patient groups typically receiving cognitive training such as patients with stroke or head traumas e are relatively comparable with regard to their cognitive profile typically with first impairments in memory and executive functions, and (iii) from an economic point of view, group trainings are more cost-effective. However, where available in the trainings' material, levels of difficulty corresponding to the group's cognitive level were used. The CG did not receive any training between test sessions (Fig. 1).

70 PD patients were included, 5 were excluded from analyses (NV/MF/CG: n ¼ 3/2/0) because they participated in less than 75% of the training sessions (reasons [NV/MF/CG]: hospital stay [n ¼ 2/2/0], changed medication [n ¼ 1/0/0]). 65 patients were included in the final analyses. Neither the three groups significantly differed in age, education, overall cognitive, affective state at baseline (Table 1), nor did the group of dropped out patients from the group of analyzed patients. At baseline, we classified all patients according to the MDS task force Level I guidelines for defining PD-MCI [3]. 50 patients were classified as cognitively unimpaired, 15 patients fulfilled criteria for PD-MCI (NV/MF/CG: n ¼ 5/6/4). Relevant (at least mild) depression scores (BDI-II>13) were found for 13 patients (NV/MF/CG: n[raw scores] ¼ 5[14,15,18,19,27]/4[15,15,20,20]/4[15,16,21,45]). H&Y

2.3. Inclusion and exclusion criteria and clinical examinations 70 patients with idiopathic PD according to UK Brain Bank Criteria [14] were recruited. Exclusion criteria were suspected dementia (Mini Mental State Examination, MMSE<25), other neurological or psychiatric diseases (except for depression), impaired hearing or sight, and treatment with DBS. Only patients who completed at least 75% (>9 sessions) of the training were included in the analysis. Change of medication from pre- to posttest served as a post-hoc exclusion criterion. 2.4. Clinical examination

1198

A. Petrelli et al. / Parkinsonism and Related Disorders 20 (2014) 1196e1202

Fig. 1. Study design.

stages ranged between I and III, UPDRS motor score between 5 and 39 (Table 1). PD subtypes classified according to a procedure described by Eggers et al. [18] (2011) are indicated in Table 1. Total dopaminergic treatment was calculated as the levodopa equivalent daily dose (LEDD) according to Tomlinson et al. (2010; Table 1) [19].

3. Results 3.1. CT versus CG: testing hypothesis I When the treatment groups were compared to the CG (Tables 1 and 2) significant time  treatment interactions were found: for the primary outcome measures only the NV-group improved in working memory (DemTect, digit span reverse: F ¼ 5.36, p ¼ .03 [parts2 ¼ .12; medium]) and short-term memory (Memo, word list learning: F ¼ 8.86, p ¼ .00 [parts2 ¼ .19; large]; trend in DemTect, word list learning: F ¼ 3.13, p ¼ .09). MF patients did not significantly benefit in any cognitive score compared to CG, although there were trends in verbal long-term memory (DemTect, delayed recall: F ¼ 3.74, p ¼ .06) and executive functions (verbal fluency of DemTect: F ¼ 3.01, p ¼ .09). For the secondary outcome measures, only MF patients benefited in depression scores (BDI-II: F ¼ 6.09, p ¼ .02 [parts2 ¼ .18; large]). Importantly, a re-analysis without the person from the CG with depression (BDI-II ¼ 45) did not yield different results (p ¼ .03). 3.2. Structured versus unstructured CT: testing hypothesis II When treatment conditions were compared, one significant time  treatment interaction was found for working memory (DemTect: F ¼ 6.75, p ¼ .01 [parts2 ¼ .15; large effect], NV > MF). Patients in the NV-group showed a tendency to benefit more in verbal short-term memory (DemTect: F ¼ 3.32, p ¼ .08). 4. Discussion The main findings of our prospective RCT are that (i) relevant cognitive functions e i.e. memory (short-term) and executive functions (working memory) e can be enhanced by CT in nondemented PD patients, and that (ii) the effects depend on the type of training. According to a recent review on non-pharmacological interventions in PD patients [8], only four RCTs used CT (two of

which in combination with exercise training, but all with comparable lengths of training between three and six weeks) e all showing positive effects on cognition. Thus, this study can be counted as one of the first RCTs of cognitive (and especially pure cognitive) training demonstrating benefits in PD patients. It is the first study showing differential effects of different CT in PD patients, comparing two CT with each other and additionally each training group with an inactive CG. The notion that the effects found in our study are of clinical significance comes from the facts that (i) effect sizes as indicated by partial eta-squared are medium to large and that (ii) effects were observed in working memory (and also short term memory) which clearly belong to the cognitive domains which are frequently affected in PD [3] and have a major impact on other cognitive functions [20]. Furthermore, as the tests used in the pre- and posttest to examine these functions differed from the tasks in the training, it can be assumed that the found effects were caused by an improvement of the underlying cognitive domain. In the structured training NV, PD patients improved in verbal short-term memory. This is consistent with the only other study that has examined effects on verbal short-term memory after CT so far, demonstrating a trend in the California Verbal Learning Test “list A1” (p ¼ .06) and “short-delay free recall” (p ¼ .10) [10]. The NV-group also showed significant gains in working memory (digit span reverse). Given that working memory can be assigned to the domain of executive functions [21], these results are also in line with other studies showing improvement of executive functions after CT [8]. This is of high relevance, as executive functions are frequently impaired in PD patients [8], are the domain most frequently affected in newly diagnosed PD patients, and belong to the most frequent symptoms in PD-MCI [3,21]. The fact that PD patients did not improve in verbal long-term memory is also in line with the already mentioned study by París et al. (2011) [10]. However, even if these results are consistent, they were unexpected, because strategies to enhance memory and longterm memory in particular, are one focus of the NV training. Importantly, even if patients with PD might have difficulties with strategy learning [20e23], Goebel et al. (2010) assume that they can nevertheless learn strategies if time is not limited [20]. A possibly underlying AD pathology in this group could have covered the effects of CT on memory [24]. However, it is possible that memory

Table 1 Patients characteristics (means [standard deviation] and ranges) and test scores (for covariates adjusted means [standard error] and 95% Confidence Interval) for NEUROvitalis (NV), Mentally fit (MF) and control group (CG) in the pre- and posttest. (primary outcomes are highlighted in grey). For references see Supplementary material (online). Patients characteristics

NV

MF

CG

Mean (standard deviation) range Socio-demographics

Clinical measures

N (gender; \:_) Agen.s. Educationn.s. Duration of disease (months)n.s. LEDDn.s. UPDRS motor score (part III)n.s. Tremor-dominant/akinetic-rigid/mixedn.s.

22 (12:10) 69.2 (4.9) 55e76 13.1 (3.8) 8e20 66.2 (39.5) 397.4 (365.4) 21.1 (7.6) 0/18/4

Testscores Domain

NV

MMSEn.s. DemTectn.s.

Attention

30 18

Brief test of attention

20

Verbal short-term

DemTect

20

10

Verbal long-term

Memo (mean 1e5) DemTect

10

Memo

10

Complex figure (recall) DemTect: digit span reverse Semantic: DemTect (z-score) Phonemic: F-A-S Complex figure

36

Visual long-term Working memory Verbal fluency

Visuo-construction

MF

CG

6 30

Reference data (age related normative data from healthy controls) <25 ¼ demented 0e8 ¼ suspected dementia 9e12 ¼ MCI 13e18 ¼ age appropriate

NV

MF

CG

27.9 (2.2) 25e30 14.2 (4.2) 9e18

28.9 (1.6) 25e30 15.1 (3.1) 9e18

28.6 (1.6) 25e30 15.5 (2.4) 10e18

27.6 (1.7) 24e30 15.0 (3.2) 9e18

45e54: 17.0 (2.7) 50e59: 17.3 (2.6) 55e64: 16.7 (2.9) 60e69: 15.8 (3.3) 65e74: 15.0 (3.6) 70e79: 14.5 (3.5) 75e84: 13.5 (3.3) <60: 13.1 (2.2) >60: 12.0 (2.2)

15.7 (1.0) 13.7e17.8

15.6 (0.8) 14.1e17.1

14.8 (0.8) 13.2e16.3

15.5 (0.8) 13.8e17.2

13.9 (0.6) 12.6e15.2 7e20 7.3 (0.3) 6.6e7.9 4.6 (0.5) 3.5e5.6 4.6 (0.6) 3.3e5.9 18.5 (1.3) 15.8e21.1 5.0 (0.2) 4.6e5.4 0.4 (0.2) 0.9e0.1

12.8 (0.6) 11.5e13.9 6.0 (0.3) 5.3e6.7 5.2 (0.5) 4.1e6.2 4.5 (0.6) 3.2e5.7 16.8 (1.3) 14.2e19.4 4.4 (0.2) 4.1e4.8 0.1 (0.2) 0.6e0.4

12.6 (0.7) 11.3e13.9 6e17 6.2 (0.4) 5.4e6.9 4.9 (0.5) 3.8e6.0 4.3 (0.7) 2.9e5.7 17.9 (1.4) 15.2e20.7 4.6 (0.2) 4.2e5.0 0.3 (0.3) 0.9e0.2

<60: 7.3 (0.7) >60: 6.6 (0.9) <60: 5.6 (1.9) >60: 4.1 (1.8) <60: 6.7 (1.2) >60: 6.7 (1.5) 19.4 (6.1) <60: >60: <60: >60:

4.9 (1.0) 4.8 (0.9) 23.5 (5.1) 21.9 (5.6)

e

e

35.9 (2.5) 30.8e40.9

32.2 (2.6) 26.9e37.4

35.8 (2.6) 30.5e41.1

36

34.4 (2.2)

Post-test

Mean (standard deviation) range

27.9 (1.9) 25e30 13.7 (2.8) 9e18

28.1 (1.7) 25e30 13.9 (2.6) 10e18

Adjusted mean (standard error) 95% confidence interval 14.1 (0.8) 14.7 (0.9) 12.8e16.5 12.4e15.8

11.9 (0.6) 10.6e13.2 6e18 6.2 (0.3) 5.7e6.8 3.7 (0.5) 2.6e4.7 4.3 (0.4) 3.5e5.0 15.7 (1.8) 12.1e19.4 4.2 (0.2) 3.7e4.6 0.3 (0.3) 0.8e0.2

11.4 (0.6) 10.3e12.6 5.7 (0.3) 5.1e6.3 3.9 (0.5) 3.0e4.9 3.8 (0.6) 2.6e4.9 14.3 (1.6) 11.0e17.6 4.4 (0.2) 3.9e4.8 0.5 (0.3) 1.0e0.1

12.4 (0.6) 11.1e13.7 8e19 6.3 (0.3) 5.7e6.9 4.8 (0.5) 3.7e5.8 4.3 (0.4) 3.5e5.0 14.2 (1.9) 10.4e18.0 4.5 (0.2) 4.1e4.9 0.2 (0.3) 0.7e0.4

32.0 (2.9) 26.1e37.9

33.1 (2.9) 27.2e38.9

36.6 (3.1) 30.4e42.8

A. Petrelli et al. / Parkinsonism and Related Disorders 20 (2014) 1196e1202

Max. score

Overall cognitive function

Executive functions

21 (9:12) 69.1 (11.6) 45e89 12.8 (3.3) 8e20 65.0 (52.8) 278.3 (170.2) 20.9 (6.6) 0/19/2

Pre-test Sub-domain

Cognitive outcomes

Memory

22 (7:15) 68.8 (6.7) 57e85 13.6 (3.2) 8e20 71.3 (56.2) 322.6 (155.9) 19.7 (8.5) 0/17/5

(continued on next page) 1199

27.3 (3.7) 19.9e34.8 24.9 (3.9) 16.9e32.8 23.0 (3.7) 15.6e30.5 28.9 (3.6) 21.6e36.4 25.5 (3.6) 18.1e32.9 24.5 (3.6) 17.1e31.9

8.5 (2.0) 4.4e12.6 9.2 (1.8) 5.5e12.9 11.1 (1.9) 7.1e15.1 10.6 (1.9) 6.7e14.5

a

n.s.

¼ no significant differences between groups in the pre-test. ¼ higher scores indicate higher impairment.

100 PDQ-39 total scorea Quality of life

Non-cognitive outcomes Depression

Patients characteristics

Table 1 (continued )

BDI-IIn.s,

a

3

Depression: 0e13: minimal 14e19: mild 20e28: moderate 29e63: severe Range: 0 (perfect health) e 100 (worse health)

10.0 (1.7) 6.6e13.5

34.7 (0.7) 33.2e36.1 34.1 (0.9) 32.1e36.1 32.7 (1.2) 30.3e35.1 33.9 (0.9) 32.1e35.7 33.3 (1.1) 31.0e35.6

Mean (standard deviation) range

CG MF NV

11.4 (2.1) 7.1e15.7

A. Petrelli et al. / Parkinsonism and Related Disorders 20 (2014) 1196e1202

33.1 (1.0) 30.9e35.2

1200

strategies might be effective if they are trained intensely enough to become automated and to be applied in everyday life as well as the test situation. París et al. (2011) found a significant improvement in visual long-term memory [10], while no effects were found in the present study. Importantly, their training program included focused training of visuo-spatial abilities with a computer training, while NV has few non-verbal memory training elements. Thus, this finding further corroborates the notion [25] that training effects usually refer to the trained domains with no or rare transfer effects. Further studies should identify methods and tasks which lead to a transfer of training benefits in PD patients as it has been shown in healthy adults [26]. No significant cognitive gains were found in the unstructured training group (MF). Thus, our data allow the conclusion that nondemented PD patients might profit more from a structured training program in which specific cognitive domains are trained in domain-focused sessions. In one report on CT in healthy elderly individuals (structured and unstructured training versus waiting list), best effects were achieved with the structured training [27]. The structure of NV may also have helped PD patients to acquire new abilities and integrate gained knowledge with established competence. This effect might even have been enhanced by background knowledge induced by the psychoeducational elements. Only the MF-group improved significantly in depression scores (BDI-II). Thus, this unstructured approach seems to be favorable with regard to the affective state of PD patients. Possibly, in a less structured setting the focus lies on social interaction (“training the brain together”), rather than in focused sessions. Communications and group conversations are an integral part of MF, whereas there is less time for social interaction in NV. No group improved in quality of life (PDQ-39). This is consistent with previous research demonstrating no benefit after CT in PD [8]. Some limitations to the present study have to be taken into account when interpreting the results. First, with 21e22 patients per group, the sample size is rather small. However, in previous studies, sample sizes were similar or even smaller except for the study by Reuter et al. (2012; n ¼ 222) [12]. With a larger patient sample, trends for improvement as demonstrated in long-term memory and verbal fluency for the unstructured training MF might stabilize and the differences between structured and unstructured training might be modified. Secondly, no post-tests were examined with those five patients, who were excluded, so no intention-to-treat-analyses could be done. Nevertheless, these patients did not differ from the 65 analyzed patients in relevant aspects (age, education, duration of disease) at baseline, so a potential drop-out-bias should be ruled out. Third, only MDS PD-MCI Level I criteria and not Level II criteria were applied for MCI diagnosis. Fourth, no follow-up data on long-term effects are reported. Thus, while the potential of short-term cognitive enhancement by cognitive treatment can be demonstrated, the question whether or not it may prevent cognitive decline needs further research. Finally, our two training programs differed not only in structure but also in elements added to the pure cognitive tasks: NV included psychoeducation, MF included unspecific conversations. Thus, it remains speculative to which specific elements the training effects can be attributed. To conclude, all RCTs available investigating pure CT including this study show that PD patients benefit from CT in various functions (executive functions [10,11], attention [10], memory [10], information processing speed [10], visual-spatial abilities [10]) affirming that it is an effective intervention strategy. A further RCT showed an improvement of speed of processing after 20 sessions of training specifically targeting this

A. Petrelli et al. / Parkinsonism and Related Disorders 20 (2014) 1196e1202

1201

Table 2 Results of the three analyses of variance adjusted for the covariates age, education and duration of disease: NEUROvitalis (NV) vs. control group (CG), Mentally fit (MF) vs. CG and NV vs. MF (primary outcomes are highlighted in grey). Domain

Sub-domain

Interaction effects time  training

Test

Hypothesis I

Hypothesis II

NV versus CG

Cognitive outcomes Attention Memory

Attention Verbal short-term

Brief test of attention DemTect Memo DemTect Memo Complex figure (recall) DemTect: digit span reverse Semantic: DemTect (z-score) phonemic: F-A-S Complex figure (copy)

Verbal long-term

Executive functions

Visual long-term Working memory Verbal fluency

Visuo-construction Figure copy Nonecognitive outcomes Depression Quality of life

BDI-II total score PDQ-39 total score

MF versus CG

NV versus MF

F

p

Part s2

F

p

Part s2

F

p

Part s2

1.25 3.13 8.86 .97 .15 .17 5.36 .08 3.05 .10

.27 .09 .00** .33 .70 .70 .03* .79 .09 .75

.04 .08 .19b .03 .00 .01 .12a .00 .08 .00

1.01 2.28 1.78 3.74 .94 .31 .00 3.01 .04 .28

.32 .14 .19 .06 .34 .58 .97 .09 .84 .60

.03 .06 .05 .09 .03 .01 .00 .07 .00 .01

.28 .46 3.32 .17 .46 .01 6.75 2.54 2.75 .00

.60 .50 .08 .68 .50 .91 .01* .12 .11 .97

.01 .01 .08 .00 .01 .00 .15b .06 .07 .00

.34 .94

.03 .00

6.09 .16

.02* .69

.18b .01

1.35 .24

.25 .63

.04 .01

.95 .01

All significant effects should be printed in bold type. * ¼ p < .05; ** ¼ p < .01. a ¼ medium effect. b ¼ large effect.

domain [28]. Further promising results are reported in one nonrandomized controlled study, also showing improvements with medium to large effect sizes for learning and memory after a multicomponent group training (psychoeducation and CT) with 14 sessions over a seven-week period [29]. Zimmermann et al. (2014) found that PD patients benefit from a cognition-specific computer-based CT program as well as from a motioncontrolled computer sports game that is not cognition-specific (both four weeks) [30]. However, data are still far from being conclusive. Short- and long-term effects on specific cognitive as well as on non-cognitive domains including quality of life will have to be studied. It has to be analyzed whether MCI or dementia in PD patients without impairment might be delayed. Finally, the interplay of nonpharmacological and pharmacological interventions should be investigated [8]. Author's roles

Petrelli Kaesberg Barbe Timmermann Fink Kessler Kalbe

Conception & design of the study/acquisition of data/analysis & interpretation of data

Drafting the article/revising it critically for important intellectual content

Final approval of the version to be submitted

X X X X X X X

X X X X X X X

X X X X X X X

Lars Timmermann received payments as a consultant for Medtronic Inc, Boston Scientific, Bayer Healthcare, UCB Schwarz Pharma. L.T. received honoraria as a speaker on symposia sponsored by TEVA Pharma, L und BeckPharma, Bracco, Gianni PR, MedasPharma, UCB Schwarz Pharma, Desitin Pharma, Boehringer Ingelheim, GlaxoSmithKline, Eumecom, Orion Pharma, Medtronic, Boston Scientific, Cephalon, Abott, GE Medical, Archimedes, Bayer. The institution of L.T., not L.T. personally received funding by the German Research Foundation, the German Ministry of Education and Research, Manfred und Ursula Müller Stiftung, KlühStiftung, €ln Fortune, Hoffnungsbaum e. V., NBIA DISORDERS SOCIETY USA, Ko Medtronic, Deutsche Parkinson Vereinigung. Archimedes Pharma, Abott, Bayer, UCB, zur Rose Pharma, TEVA. Gereon R. Fink received honoraria for speaking engagements or educational activities from TEVA (speaker, 2009), GlaxoSmith-Kline (speaker, 2009), Boehringer Ingelheim (speaker, 2008). He received research support by the Bundesministerium für Bildung und Forschung, 10GO 0204, Deutsche Forschungsgemeinschaft (principal applicant, 2002e2007, 2010 -) and the Volkswagen-Stiftung, principal investigator, 2000- 2002. Josef Kessler is co-Author of the NEUROvitalis program. Elke Kalbe received honoraria as a speaker on symposia sponsored by Novartis Pharma GmbH, TEVA Pharma, and Merck Serono GMBH. The institution of E.K., not E.K. personally received funding by the German Ministry of Education and Research. She is also coAuthor of the NEUROvitalis program. Funding sources for study Not applicable. Acknowledgments

Financial disclosure related to research covered in this article Annette Petrelli reports no conflict of interest. Stephanie Kaesberg is co-Author of the NEUROvitalis program. Michael T. Barbe received honoraria for speaking engagements from GE Medical and Medtronic Inc. and travel grants from Medtronic Inc..

This study was in part supported by the German Ministry for Education and Research (BMBF) (LANDSCAPE study project-nr: 01GI1008c). We thank our colleagues Dipl.-Psych. Julia Rahe, MA Geront Jennifer Liesk, Dipl.-Psych. Jan Rosen (Institute of Gerontology & Center for Neuropsychological Diagnostics and Intervention [CeNDI], University of Vechta) as well as Dr. John V Hindle

1202

A. Petrelli et al. / Parkinsonism and Related Disorders 20 (2014) 1196e1202

(School of Medical Sciences, Bangor University, United Kingdom) for their very helpful comments to the manuscript. Furthermore, we especially thank the patients and support groups for their active participation. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.parkreldis.2014.08.023. References [1] Muslimovic D, Post B, Speelman JD, Schmand B. Cognitive profile of patients with newly diagnosed Parkinson disease. Neurology 2005;65:1239e45. [2] Aarsland D, Brønnick K, Fladby T. Mild cognitive impairment in Parkinson's disease. Curr Neurol Neurosci 2011;11:371e8. [3] Litvan I, Aarsland D, Adler CH, Goldman JG, Kulisevsky J, Mollenhauer B, et al. MDS task force on mild cognitive impairment in Parkinson's disease: critical review of PD-MCI. Mov Disord 2011;26:1814e24. [4] Reijnders J, van Heugten C, van Boxtel M. Cognitive interventions in healthy older adults and people with mild cognitive impairment: a systematic review. Ageing Res Rev 2013;12:263e75. [5] Leroi I, McDonald K, Pantula H, Harbishettar V. Cognitive impairment in parkinson Disease: impact on quality of life, disability, and caregiver burden. J Geriatr Psych Neur 2012;25:208e14. [6] Emre M, Aarsland D, Brown R, Burn DJ, Duyckaerts C, Mizuno Y, et al. Clinical diagnostic criteria for dementia associated with Parkinson's disease. Mov Disord 2007;22:1689e707. [7] Huckans M, Hutson L, Twamley E, Jak A, Kaye J, Storzbach D. Efficacy of cognitive rehabilitation therapies for mild cognitive impairment (MCI) in older adults: working toward a theoretical model and evidence-based interventions. Neuropsychol Rev 2013;23:63e80. [8] Hindle JV, Petrelli A, Clare L, Kalbe E. Nonpharmacological enhancement of cognitive function in Parkinson's disease: a systematic review. Mov Disord 2013;28:1034e49. [9] Murray DK, Sacheli MA, Eng JJ, Stoessl AJ. The effects of exercise on cognition in Parkinson's disease: a systematic review. Transl Neurodegener 2014;3:5. http://dx.doi.org/10.1186/2047-9158-3-5. [10] París AP, Saleta HG, de la Cruz Crespo Maraver M, Silvestre E, Freixa MG, Torrellas CP, et al. Blind randomized controlled study of the efficacy of cognitive training in Parkinson's disease. Mov Disord 2011;26:1251e8. [11] Sammer G, Reuter I, Hullmann K, Kaps M, Vaitl D. Training of executive functions in Parkison's disease. J Neurol Sci 2006;248:115e9. [12] Reuter I, Mehnert S, Sammer G, Oechsner M, Engelhardt M. Efficacy of a multimodal cognitive rehabilitation including psychomotor and endurance training in Parkinson's disease [serial online] J Aging Res 2012;2012:235765.

[13] Baller G, Kalbe E, Kaesberg S, Kessler J. NeuroVitalis. Ein neuropsychologisches € rderung der geistigen Leistungsfa €higkeit. Ko €ln: Gruppenprogramm zur Fo Prolog; 2010. [14] Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry 1992;55:181e4. [15] Hoehn MM, Yahr MD. Parkinsonism: onset, progression and mortality. Neurology 1967;17:427e42. [16] Fahn S, Elton RL. The unified Parkinson's disease rating scale, vol. 2. Florham Park (NJ): McMillan Healthcare Information; 1987. [17] Yaffe K, Hoang T. Nonpharmacological treatment and prevention strategies for dementia. Contin (Minneap Minn) 2013;19:372e81. [18] Eggers C, Kahraman D, Fink GR, Schmidt M, Timmermann L. Akinetic-rigid and tremor-dominant Parkinson's disease patients show different patterns of FPCIT single photon emission computed tomography. Mov Disord 2011;26: 416e23. [19] Tomlinson CL, Stowe R, Patel S, Rick C, Gray R, Clarke CE. Systematic review of levodopa dose equivalency reporting in Parkinson's disease. Mov Disord 2010;25:2649e53. [20] Goebel S, Mehdorn HM, Leplow B. Strategy instruction in Parkinson's disease: influence on cognitive performance. Neuropsychologia 2010;48:574e80. [21] Kudlicka A, Clare L, Hindle JV. Executive functions in Parkinson's disease: systematic review and meta-analysis. Mov Disord 2011;26:2305e15. [22] Smith SJ, Souchay C, Conway MA. Improving metamemory in ageing and Parkinson's disease. Age Aging 2010;39:116e9. [23] Moro dos Santos C, BisolBalardin J, QuartiIrigaray T, Schroder N, Rieder CRM, Bromberg E. Incidental encoding strategies did not improve contextual memory in Parkinson's disease patients. Neurorehabil Neural Repair 2009;24:450e6. [24] Irwin DJ, Lee VM, Trojanowski JQ. Parkinson's disease dementia: convergence of a-synuclein, tau and amyloid-b pathologies. Nat Rev Neurosci 2013;14: 626e36. [25] Ellfolk U, Huurinainen S, Joutsa J, Karrasch M. The effect of encoding condition on free recall in Parkinson's disease: incidental and intentional memory are equally affected. Clin Neuropsychol 2012;26:909e25. [26] Buitenweg JIV, Murre JMJ, Ridderinkhof KR. Brain training in progress: a review of trainability in healthy seniors. Front Hum Neurosci 2012;6. http:// dx.doi.org/10.3389/fnhum.2012.00183. [27] Jaeggi SM, Buschkuehl M, Jonides J, Perrig WJ. Improving fluid intelligence with training on working memory. Proc Natl Acad Sci 2008;105:6829e33. s EG, Zesiewicz TA, Uc EY. Ran[28] Edwards JD, Hauser RA, O'Connor ML, Valde domized trial of cognitive speed of processing training in Parkinson disease. Neurol 2013;8:1284e90. [29] Naismith SL, Mowszowski L, Diamond K, Lewis SJ. Improving memory in Parkinson's disease: a healthy brain ageing cognitive training program. Mov Disord 2013;28:1097e103. [30] Zimmermann R, Gschwandtner U, Benz N, Hatz F, Schindler C, Taub E, et al. Cognitive training in Parkinson disease: cognition-specific vs nonspecific computer training. Neurol 2014;82:1219e26.