Parkinson’s disease (PD) is associated with a decline in postural control, walking ability, an increased risk for falling, and a deteriorating quality of life (1–3). Although pharmacological treatments are still the mainstream to treat motor symptoms, physical exercise can also improve PD patients’ postural control and mobility (2–6). Recommendations urge the use of high-intensity exercise stimulus to produce rapid and lasting improvements in PD symptoms (4,7,8). However, the results are inconsistent. Even at the same disease stage, treadmill exercise improved gait speed and cardiovascular fitness independent of exercise intensity (9). In addition, high compared with low frequency exercise can unfavorably affect functional outcomes (10). Yet there is also evidence that exposing PD patients to high-intensity exercising training can functionally meaningfully improve early-stage PD patients’ symptoms (8). How long such exercise effects last after the exercise stimulus is withdrawn is also unclear. Despite recommendations to measure the effects at least for 24 months, in 16 studies the average follow-up time was 5.5 months (2,5). The only study with a 24-month-long maintenance program reported favorable effects on selected motor symptoms but also numerically almost identical elbow flexion torque at baseline (50.8 N·m) and at 24 months (50.2 N·m) (11). Yet a second booster dose of multidisciplinary intensive rehabilitation exercise at year 1 after an initial bout at baseline even without a maintenance program, improved, at year 2, UPDRS III scores, 6-min walking distance, and timed up and go performance by 3.4 points, 41.1 m, and 1.1 s, respectively (12).
Whether exercise can reduce patients’ drug dose is unclear. In one case, l-dopa equivalent increased moderately by 38.4% in the intensive exercise group compared with 327.4% in the no-exercise controls at the end of the 2-yr follow-up period, suggesting that intense exercise without a maintenance program could moderate drug dose (12). Despite the 2-yr low-intensity exercise maintenance program, l-dopa equivalent, however, still increased by 29% (11). It is thus unclear if a high-intensity and long-term exercise maintenance program could reduce the increase normally seen in PD patients’ medication.
The purpose of the present study was to determine the immediate and long-term effects of a 3-wk-long, high-intensity and high-frequency agility program on PD patients’ motor and clinical symptoms with and without a 2-yr-long high-intensity agility maintenance program. We expected that patients would tolerate the short-term 3-wk initial high-intensity agility program and motor and nonmotor symptoms would improve and that the maintenance program would sustain these improvements and slow the progression of symptoms. Based on previous studies (11,12), we also expected that the maintenance program would slow the increase in l-dopa equivalent levels.
Design and patients
This is a three-group, randomized clinical trial involving PD patients who met the UK Brain Bank criteria and were of stages 2 to 3 on the Hoehn and Yahr scale. Figure 1 shows the Consort diagram. Form the hospital database and the outpatient clinic we identified 72 patients who met inclusion criteria based on medical records. Of these, 17 were excluded and the remaining 55 randomized into: exercise+maintenance (E + M, n = 19, 11 M); Exercise only group (E, n = 16, 6 M), and to a no exercise and no maintenance control group (C, n = 20, 12 M; Table 1). At the time of the start of the study and for a 2-yr period preceding it, none of the patients were enrolled in rehabilitation. The initial high-intensity and high-frequency agility E program lasted 3 wk. The M program lasted 2 yr. All patients were assessed eight times: before and after the 3-wk exercise program and at 3, 6, 9, 12, 18, and 24 months. Wait-listed patients in C had the opportunity to enroll in the exercise program after the end of the trial. After the 3-wk initial exercise intervention, patients in E and C were not enrolled in an exercise or maintenance program during the 2-yr period.
Patients were recruited from the hospital database. An initial screening established disease severity by the language-validated version of Movement Disorder Society Unified Parkinson Disease Rating Scale, Motor Experiences of Daily Living (MDS-UPDRS-M-EDL). A preliminary screening included a full neurological exam and a mobility evaluation. The exam ensured that all included patients had mobility difficulty and postural instability based on a qualitative assessment of gait and postural stability, turns, rigidity, inter-joint coordination, trunk posture, and equilibrium while subjects walked forward, backwards, and sideways. In a separate visit, a neuropsychologist evaluated patients’ cognitive function. Patients were excluded with brain abnormalities based on a diagnostic MRI, Mini Mental State Examination score < 24, a Beck Depression Inventory score > 40, severe cardiac disease, uncontrolled diabetes, a history of stroke, traumatic brain injury, seizure disorder, past or current deep brain stimulation, or current participation in a self-directed or formal group exercise program. All patients remained “on” medication so that the assessments at baseline and after the intervention and each exercise session occurred 1-2 h after patients took PD medications.
The principal investigator performed the randomization. He drew a colored ribbon from a covered box and attached one ribbon to each patient’s folder (E, red; E + M, blue; C, green). Two physical therapists and a physical therapy assistant administering the tests, were masked to patients’ group assignments. In the familiarization session patients practiced each test and watched the Xbox kinect programs, a key element of the intervention. Patients gave written informed consent to participate in the study. The University Hospital’s Ethics Committee (IKEB) approved the study protocol. The trial was registered at Clinicaltrials.gov (NCT03193489).
The primary outcome was MDS-UPDRS-M-EDL, which is sensitive to changes in a broad spectrum of PD symptoms (13). We accepted changes >3.1 points as a minimal clinically important difference (14). The lead physical therapist administered this test in person every time to every patient to assess motor signs of PD.
Secondary outcomes measured changes in health-related quality of life using: 1) Schwab and England Activities of Daily Living Scale (ADL); 2) EuroQol EuroQol five dimensions questionnaire., and 3) the Parkinson’s Disease Questionnaire (PDQ-39, minimal clinically important difference: 4.7 points) (15). The Beck Depression Inventory measured depression and the timed up and go test (TUG) quantified mobility. We quantified postural stability by the magnitude of sway measured on a force platform while standing in a wide and a narrow stance with eyes open or closed. Participants stood for 20 s in each of the four conditions administered in order of: 1) eyes open wide and 2) narrow stance followed by standing with eyes closed 3) in a wide and 4) in a narrow stance. The outcome was the 3D path of the center of pressure (COP) (in mm). The testing order was standardized among patients and testing sessions. Adverse events were not systematically assessed.
The Exercise program comprised a high-intensity agility intervention, detailed previously in the supplementary material of that article comprising a different group of patient (16). Briefly, E + M and E completed fifteen 1-h-long sessions over 3 wk and targeted deficits in postural control and mobility. Three therapists delivered the program by having patients exercise in small groups at individual times only in the hospital’s physical therapy gym. Therapists demonstrated most exercises, mingled among patients on the exercise floor to closely supervise and spot them for safety. Patients were asked not to enroll in any other activity programs and perform additional exercises at home on their own. Patients exercised without shoes on a 26-mm-thick Theraband-carpeted floor. After 10 min of warm-up, patients completed a 20-min block of sensorimotor and visuomotor agility training and a 20-min block of sensorimotor agility training using the X-box virtual reality exergame (Microsoft xbox 360 core system with kinect, Microsoft Corp.) (17). Each session ended with 10 min of cool down. The sensorimotor and visuomotor agility training included: 1) gait training; 2) coordination training; 3) posture training with and without an augmented sensory input; 4) balance exercises with and without a peer, assistive devices, height stimuli, surface modifications, postural changes, shifts between tasks, and directional changes; 5) body scheme exercises; and 6) posture-corrective exercises. We detailed previously exercise dosing, surface manipulations, task numbers, task types, feedback, and other methods to increase and manipulate motor and sensory stimuli, including the sophisticated use of the X-box virtual reality exergame and how patients kept an exercise log to record symptoms, fatigue, and attendance. A video clip in Supplement 1 shows patients exercising (see Video, Supplemental Digital Content 1, PD patients performing exergaming agility exercises, https://links.lww.com/MSS/B403). The nonexergaming and exergaming each represented about 50% of the total exercise time. The average heart rate and rate of perceived exertion was 120.6 bpm and 13.6 or about 80% age-predicted maximum heart rate and ”somewhat hard/hard” on the 20-point Borg scale (unpublished data).
After the 3-wk-long, daily, high-intensity exercise intervention, E + M continued the Maintenance program three times per week for 2 yr in the hospital’s physical therapy gym using the same exercises used in the 3-wk-long initial exercise program. The three therapists supervised each session attended by small groups of three to five patients who exercised at the same time of the day for 1 h. The aim of the maintenance program was to determine if patients can endure a high-intensity rehabilitation program for an extended time period and if such a program can slow disease progression. E did not perform the maintenance phase and C received no exercise therapy and no maintenance either.
We estimated the number of participants needed for a significant group (E + M, E, C) by time (0 and 3 wk, 3, 6, 9, 12, 18, 24 months) analysis of variance with repeated measures on Time for a change of four points caused by the initial intense intervention (>3.1 functionally meaningful change) (14). Using an alpha of 0.05, 1-beta (power) of 0.8, three groups, a correlation of 0.5 between repeated measures, the total sample size needed was 49 patients. Anticipating dropouts, we randomized 55 patients.
Data are expressed as mean ± SD. The variables were normally distributed based on the Shapiro–Wilk test. The main analysis was a Group (E + M, E, C) by Time (0 and 3 wk, 3, 6, 9, 12, 18, 24 months) analysis of variance with repeated measures on Time. In case of an interaction, we used a Tukey post hoc contrast to determine the means that differed at P < 0.05. We also compared at baseline those nine patients who deceased over the 2 yr with those who completed the trial. We computed Pearson correlations between changes in the primary and secondary outcomes to explore potential mechanistic links underlying improvements in patients’ mobility and clinical symptoms. The level of significance was set at P < 0.05. All statistical analyses were conducted with SPSS version 22.
Table 1 shows that the groups were similar at baseline. During the 3-wk high-frequency exercise program and also during the 2-yr-long maintenance program, attendance and compliance were 100%, dropout was 0%, and there were no adverse events, which were not assessed systematically.
The 3-wk-long agility program improved MDS-UPDRS M-EDL significantly (P < 0.05) but similarly by 30.4% (±10.23) or 6.3 points (±3.06) in E + M and by 42.8% (±9.43) or 7.8 (±1.57) points in E. These changes were greater than the nonsignificant changes in C (group–time interaction, F12,258 = 32.7, P = 0.001, Table 2, Fig. 2A).
E + M sustained the exercise-induced benefits. In E, the exercise-induced improvements were still present at 3 months. C exhibited a gradual worsening over the 2 yr. At year 2, there was a 12.4 points difference in favor of E + M versus C (P < 0.05). Over 2 yr, the MDS-UPDRS M-EDL score had decreased by six points in C.
The group–time interaction for l-dopa equivalents was not significant (P = 0.662) and the dose increased by 97.4 mg·d−1 or 11.4% in the three groups combined (Time main effect), F4,168 = 3.6, P = 0.008 (Fig. 3.)
The agility program improved the PDQ by 26.0% (±7.36) in E + M and by 28.9% (±9.31) in E, more than the 6.8% (±16.85) worsening in C (interaction, F12,258 = 9.9, P = 0.001, Table 2, Fig. 2B). E + M kept the exercise-induced improvements in PDQ for 2 yr at a steady level. In E, the exercise effects were still present at 12 month (Fig. 2B). At 24 months, E + M versus E and E + M versus C had 15.3 and 24.4 points better PDQ score (both P < 0.05). E still had a 9.1 better score than C (P < 0.05). Over the 2 yr, the PDQ score had decreased by 20 points in C (Table 2).
The exercise intervention improved the Beck Depression Index (F12,258 = 12.5), the Schwab and England ADL inventory (F12,258 = 8.9), the EQol visual analogue scale (VAS) scores (F12,258 = 10.3), and the EQoL summed scores (F12,258 = 21.5) in E + M (range of improvements: 13% to 21%, all P ≤ 0.001) and in E (14% to 20%, all P < 0.05, Table 2). In E, these effects lasted for 3 months. At 24 months, E + M still showed the exercise-induced gains and E returned to baseline. Compared with E + M at 24 months, the scores in C were all worse in the Beck Depression Index, Schwab and England ADL inventory, the EQol VAS scores and in the EQoL summed scores (all P < 0.05).
Timed up and go test improved by 6.3 s (±2.75) in E + M and by 6.0 s (±2.96) E (all P < 0.05) compared with the 0.6 s (±0.76) in C (n.s.) (interaction, F12,258 = 20.2, P < 0.001, Table 2, Fig. 2C). These effects lasted for 18 months in E (Fig. 2C). At 24 months, E + M had 6.8 s shorter TUG time than C (P < 0.05). TUG remained unchanged over 2 yr in C (n.s.). Exercise decreased COP path in the four conditions similarly in E + M and E (range: 2.0 to 6.9 mm) and E + M sustained the exercise-induced improvements. In E, the exercise effects lasted until month 12 in the four posturography measures. At 24 months, E + M versus E had 4.7 to 2.5 mm shorter COP path in the four measures (P < 0.05), and these differences between E + M versus C had even larger (range: 4.2 to 6.7 mm, P < 0.05).
MDS-UPDRS M-EDL at baseline correlated with the change in MDS-UPDRS M-EDL at 3 wk (r = −0.803) and this correlation essentially remained unchanged by 24 months (r = −0.683, n = 18, P < 0.05). Because the primary outcome reached a plateau at month 3 during follow-up in E + M (n = 18, Fig. 2A), we determined the relationship between changes in the primary outcome, MDS-UPDRS M-EDL, for the period from baseline to 3 month and the changes over the same period in PDQ (r = 0.422), Beck depression score (r = 0.198), EQ VAS (r = −0.181), TUG (r = 0.126), and the four postural measures (range of r = 0.092 to 0.297). None of these correlations were significant (P > 0.05). The correlation between changes in MDS-UPDRS M-EDL and number of PD years was also low (r = 0.271).
Characteristics of deceased patients
One, three, and five patients, respectively, died in E + M, E, and C, with 46 of 55 original patients completing the 2-yr study. Causes of death were heart attack (n = 2), unknown (n = 3), tumor (n = 2), and stroke (n = 2), all unrelated to study. Table 3 shows the baseline comparisons between patients who died and those who were alive at the end of the 2-yr program. At baseline, there were differences between these two groups in MDS-UPDRS M-EDL and TUG.
High-intensity and high-frequency supervised sensorimotor agility exercise (3 wk, 15 sessions) improved PD patients’ motor and nonmotor symptoms. The subsequent 2-yr-long supervised maintenance program sustained but did not further improve the benefits produced by the initial 3-wk program in the eight outcomes. The favorable effects of the 3-wk agility program without the maintenance program on motor and nonmotor symptoms lasted for 3 to 12 months. Patients in the no-intervention control group declined steadily in all outcomes over 2 yr. Exercise therapy with and without the maintenance program did not reduce drug dose.
Acute exercise effects
The data contribute to the emerging picture that a variety of motor interventions can improve PD patients’ motor and clinical symptoms (2–5). The ~7.0 points (n = 35, effect size 1.2), over twice the 3.1 points of clinically meaningful improvements (14) in MDS-UPDRS M-EDL are similar to the changes of 7.3 in UPDRS III following a 4-wk-long multidisciplinary intensive rehabilitation treatment (12). The improvements correlated strongly with the baseline scores, suggesting that the intervention was particularly effective and, as hypothesized, not harmful in patients with low initial scores. Thus, high-intensity and challenging exercise therapy is effective for PD patients with a Hoehn–Yahr stage 1.2 (12) but also for patients at stage of 2 to 3 (present study, Table 1) to improve perceived and measured mobility, posture, and clinical symptoms. Future studies will determine whether or not high intensity and frequency are prerequisites to induce such acute effects on MDS-UPDRS M-EDL, as lower-intensity yoga, dance, and balance training are also effective (2–5) and superior to very low-intensity physical and occupational therapy (18).
The 3-wk-long intervention uniformly improved secondary outcomes of perceived and objectively measured functions by 13% to 55% (effect sizes: 0.53–2.54, Table 2). Because depression affects quality of life most, it was important to see that exercise improved QoL and the Beck Depression Index (3.3 points). Thus, agility training in addition to aerobic exercise can also improve PD patients’ depression (19,20). Changes in the Schwab and England (10 points), TUG (6.1 s), PDQ (13.3 points), and posturography scores suggest improved static and dynamic balance and nonmotor symptoms, confirming and for the most part exceeding changes reported previously (2–5) The high response rate in all outcomes is probably related to the suitability of the exercise stimulus, as patients attended all sessions and none dropped out.
A 2-yr-long agility maintenance program slowed the progression of PD symptoms (Fig. 2, Table 2). The maintenance program clinically meaningfully (14) further improved the primary outcome by 3.5 points at month 3 but thereafter this improved level remained unchanged. The favorable initial rapid adaptations to the 3-wk program disappeared in E so that at month 6, there were no differences (2 points, n.s.) in MDS-UPDRS M-EDL scores between E and C (Fig. 3, Table 2). The maintenance program did not further increase the gains produced by the initial intense exercise phase in the secondary outcomes but the maintenance program was necessary to sustain the initial gains in all outcomes. The data provide evidence that even short-term exercise programs can moderate PD patients’ motor and nonmotor symptoms. However, such changes are transient and for lasting neuroprotective and restorative effects to occur, PD patients need to participate in long-term maintenance programs (1–5,21).
Agility and resistance training can both improve motor and nonmotor symptoms and maintain such improvements (11,12,22). The difference between our agility and other agility and resistance training programs could be in effectiveness. In our patients the MDS-UPDRS M-EDL scores were 12.4 points lower (better) than control (Fig. 2, Table 2) in contrast to the 2.2-point difference reported at month 24 in favor of the multidisciplinary intensive rehabilitation treatment versus control (12). In this study patients’ disease severity was lower (mean Hoehn–Yahr stage of 1.2) (12) than in the present study (range Hoehn–Yahr stage of 2–3). The on-medication MDS-UPDRS III scores at month 24 after resistance training maintenance program changed little (11). Taken together, it may be necessary to keep exercise intensity high for a prolonged period to slow the progression of PD symptoms and improve MDS-UPDRS M-EDL scores by 12.4-point (Table 2). The agility program did not affect drug dose, which, against our expectation of a relative reduction, increased by 11% (12).
Most PD patients with a diagnosis of stages 2 to 3 on the Hoehn-Yahr scale present with multiple comorbidities. Those who died compared with those who completed the program differed (P < 0.05) at baseline only in two variables (MDS-UPDS M-EDL, worse score by 7.2 points; TUG, 2.9 s longer; Table 3). The suggestion emerging from these data requires confirmation as to which variables could be used to predict progression of PD symptoms in stages 2 to 3 PD patients.
The mechanisms of how a prolonged and high-intensity exercise incorporating sensorimotor and visuomotor stimuli might slow the progression of disease in PD patients remain unclear. Short-term intensive balance training challenges postural stability and produced correlated morphometric changes in gray matter of brain areas and balance behavior (23). Motor-cognitive training decreased PD patients’ reliance on frontal brain structures, resulting in improved functioning (24). At the cellular level, animal and human PET data suggest that exercise can improve dopamine signaling, leading to task-specific improvements in postural control (25–27). Such improvements in motor function are accompanied by neuroplastic changes, including improved dopaminergic signaling through an increase in striatal dopamine release, reduced dopamine reuptake, and an elevated dopamine-D2 receptor expression measured at protein and transcript levels. Sustained exercise activates neurotrophic factors, which produce anti-inflammatory and pro-regenerative effects on motor and cognition function in old adults with and without a degenerative condition (28–31). In particular, there is emerging evidence suggesting that rapid reactive movements to external and internal perturbations on unstable surfaces, as done in the present agility training study, could increase the descending neural drive leading to correlated improvements in clinical symptoms and in the magnitude, timing and rate of torque generation (32).
Without a maintenance-only group we cannot tell if the initial 3-wk-long exercise period enhanced the maintenance effects. It is likely that some of the maintenance effects were due to the attention and social contact patients received over the 2 yr in contrast to a lack of attention and contact in the E and C groups. The correlations between changes in the outcomes did not reach significance, making causation among variables not possible. As the outcomes were purely behavioral, we could not examine any potential mechanisms. To achieve and maintain the high exercise intensity, adherence, and compliance, three therapists and a designated facility were needed, conditions that may not be available in many settings. Finally, without a high-intensity comparison group such as intensive cycling (33,34), in which interaction with unstable surfaces and rapid responses to external and internal perturbations are absent, we cannot tell if in the present and past studies (4,8,12,32,35) the agility or the fitness stimulus did in fact produce the disease-slowing postural and mobility improvements, an issue we are addressing in our ongoing studies. A lack of systematic assessment of adverse events is a limitation but anecdotally and based on patients’ exercise diaries we found no evidence for program-related falls in and outside the gym. The deaths for which pathology reports were available were caused by serious medical conditions unrelated to the intervention. In conclusion, a high-intensity sensorimotor agility program with but not without a 2-yr maintenance program slowed the progression of PD patients’ motor and nonmotor symptoms without reducing drug dose.
Supported in part by the Department of Neurology, Somogy County Moricz Kaposi General Hospital, and by a `Regional Health Development’ award from the Doctoral School of the Faculty of Health Sciences, University of Pécs.
The authors declare no conflict of interest. The authors state that the results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.
The present study does not constitute endorsement by ACSM.
1. Abbruzzese G, Marchese R, Avanzino L, Pelosin E. Rehabilitation for Parkinson’s disease: current outlook and future challenges. Parkinsonism Relat Disord
. 2016;22(1 Suppl):S60–4.
2. Bloem BR, de Vries NM, Ebersbach G. Nonpharmacological treatments for patients with Parkinson’s disease. Mov Disord
3. van der Kolk NM, King LA. Effects of exercise on mobility in people with Parkinson’s disease. Mov Disord
4. Frazzitta G, Balbi P, Maestri R, Bertotti G, Boveri N, Pezzoli G. The beneficial role of intensive exercise on Parkinson disease progression. Am J Phys Med Rehabil
5. Klamroth S, Steib S, Devan S, Pfeifer K. Effects of exercise therapy on postural instability in Parkinson disease: a meta-analysis. J Neurol Phys Ther
6. Kwakkel G, de Goede CJ, van Wegen EE. Impact of physical therapy for Parkinson’s disease: a critical review of the literature. Parkinsonism Relat Disord
. 2007;13(3 Suppl):S478–87.
7. Ahlskog JE. Does vigorous exercise have a neuroprotective effect in Parkinson disease? Neurology
8. Conradsson D, Lofgren N, Nero H, et al. The effects of highly challenging balance training in elderly with Parkinson’s disease: a randomized controlled trial. Neurorehabil Neural Repair
9. Shulman LM, Katzel LI, Ivey FM, et al. Randomized clinical trial of 3 types of physical exercise for patients with Parkinson disease. JAMA Neurol
10. Pelosin E, Avanzino L, Barella R, et al. Treadmill training frequency influences walking improvement in subjects with Parkinson’s disease: a randomized pilot study. Eur J Phys Rehabil Med
11. Corcos DM, Robichaud JA, David FJ, et al. A two-year randomized controlled trial of progressive resistance exercise for Parkinson’s disease. Mov Disord
12. Frazzitta G, Maestri R, Bertotti G, et al. Intensive rehabilitation treatment in early Parkinson’s disease: a randomized pilot study with a 2-year follow-up. Neurorehabil Neural Repair
13. Horvath K, Aschermann Z, Acs P, et al. Validation of the Hungarian MDS-UPDRS: why do we need a new Parkinson scale? Ideggyogy Sz
14. Horvath K, Aschermann Z, Kovacs M, et al. Minimal clinically important differences for the experiences of daily living parts of movement disorder society-sponsored unified Parkinson’s disease rating scale. Mov Disord
15. Horvath K, Aschermann Z, Kovacs M, et al. Changes in quality of life
in Parkinson’s disease: how large must they be to be relevant? Neuroepidemiology
16. Tollár J, Nagy F, Kovács N, Hortobágyi T. A high-intensity multi-component agility intervention improves Parkinson’s patients’ clinical and motor symptoms. Arch Phys Med Rehabil
17. Pompeu JE, Arduini LA, Botelho AR, et al. Feasibility, safety and outcomes of playing kinect adventures! For people with Parkinson’s disease: a pilot study. Physiotherapy
18. Clarke CE, Patel S, Ives N, et al. Clinical effectiveness and cost-effectiveness of physiotherapy and occupational therapy versus no therapy in mild to moderate Parkinson’s disease: a large pragmatic randomised controlled trial (pd rehab). Health Technol Assess
19. Wu PL, Lee M, Huang TT. Effectiveness of physical activity on patients with depression and Parkinson’s disease: a systematic review. PLoS One
20. Adamson BC, Ensari I, Motl RW. Effect of exercise on depressive symptoms in adults with neurologic disorders: a systematic review and meta-analysis. Arch Phys Med Rehabil
21. Tomlinson CL, Herd CP, Clarke CE, et al. Physiotherapy for Parkinson’s disease: a comparison of techniques. Cochrane Database Syst Rev
22. Prodoehl J, Rafferty MR, David FJ, et al. Two-year exercise program improves physical function in Parkinson’s disease: the PRET-PD randomized clinical trial. Neurorehabil Neural Repair
23. Sehm B, Taubert M, Conde V, et al. Structural brain plasticity in Parkinson’s disease induced by balance training. Neurobiol Aging
24. Maidan I, Rosenberg-Katz K, Jacob Y, Giladi N, Hausdorff JM, Mirelman A. Disparate effects of training on brain activation in Parkinson disease. Neurology
25. Petzinger GM, Walsh JP, Akopian G, et al. Effects of treadmill exercise on dopaminergic transmission in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned mouse model of basal ganglia injury. J Neurosci
26. Toy WA, Petzinger GM, Leyshon BJ, et al. Treadmill exercise reverses dendritic spine loss in direct and indirect striatal medium spiny neurons in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (mptp) mouse model of Parkinson’s disease. Neurobiol Dis
27. Fisher BE, Li Q, Nacca A, et al. Treadmill exercise elevates striatal dopamine d2 receptor binding potential in patients with early Parkinson’s disease. Neuroreport
28. Frazzitta G, Maestri R, Ghilardi MF, et al. Intensive rehabilitation increases BDNF serum levels in parkinsonian patients: a randomized study. Neurorehabil Neural Repair
29. Alkadhi KA. Exercise as a positive modulator of brain function. Mol Neurobiol
30. Cobianchi S, Arbat-Plana A, Lopez-Alvarez VM, Navarro X. Neuroprotective effects of exercise treatments after injury: the dual role of neurotrophic factors. Curr Neuropharmacol
31. Tajiri N, Yasuhara T, Shingo T, et al. Exercise exerts neuroprotective effects on Parkinson’s disease model of rats. Brain Res
32. Silva-Batista C, Corcos DM, Barroso R, et al. Instability resistance training improves neuromuscular outcome in Parkinson’s disease. Med Sci Sports Exerc
33. Nadeau A, Lungu O, Duchesne C, et al. A 12-week cycling training regimen improves gait and executive functions concomitantly in people with Parkinson’s disease. Front Hum Neurosci
34. Uygur M, Bellumori M, Knight CA. Effects of a low-resistance, interval bicycling intervention in Parkinson’s disease. Physiother Theory Pract
35. Morberg BM, Jensen J, Bode M, Wermuth L. The impact of high intensity physical training on motor and non-motor symptoms in patients with Parkinson’s disease (pip): a preliminary study. NeuroRehabilitation