The Effects of Multimodal Exercise on Sleep Quality and Architecture, Motor Function, Cognition, Fatigue, and Systemic Inflammation in Corticobasal Syndrome: A Case Report

ABSTRACT

Importance

Corticobasal syndrome (CBS) is a rare tauopathy, with a complex pathophysiology that usually includes neuroinflammation. Parkinsonism, cognitive impairments, and sleep disturbances are common in CBS, although alterations in sleep architecture remain poorly characterized. Regular exercise has been recommended in CBS to manage gait dysfunction, balance issues, and cognitive decline. However, the effects of regular exercise on sleep quality, sleep architecture, and systemic inflammation in CBS remain unclear.

Objective

The purpose of this study was to describe the effects of regular exercise in CBS.

Design

The design of this study was a case report.

Settings

This study was conducted in an academic laboratory.

Participants

An individual with CBS participated in this case study.

Intervention

The participant completed a 12-week multimodal training program.

Main Outcome(s) and Measure(s)

Cardiorespiratory fitness level was assessed with a symptom-limited cardiopulmonary exercise test and strength with a submaximal 1-repetition maximum test. Subjective and objective sleep quality were assessed using the Parkinson Disease (PD) Sleep Scale-2 and actigraphy, respectively. Sleep architecture was evaluated with polysomnography. Cognition and motor function were assessed with the Scale for Outcomes in PD-Cognition and Movement Disorder Society-Unified Parkinson’s Disease Rating Scale part III (MDS-UPDRS-III), respectively, functional mobility with the Time Up and Go (TUG), and fatigue with the PD Fatigue Scale. Concentrations of inflammatory markers, including interleukin (IL)-1β, IL-6, IL-10, tumor necrosis factor, and C-reactive protein, were measured from serum collected after a 12-hour fasting period.

Results

Following the training program (34 sessions; 25.35 hours), improvements in fitness, objective sleep quality and architecture, cognition, TUG, and a reduction in systemic inflammation were observed. Conversely, MDS-UPDRS-III scores deteriorated, and the participant reported diminished subjective sleep quality and increased fatigue.

Conclusions

These results, which should be interpreted with caution, suggest that various clinical outcomes improved following multimodal training. Controlled studies are warranted to confirm these observations.

Relevance

This is the first case report describing the effects of a training program on sleep architecture and systemic inflammation in CBS.

INTRODUCTION

Corticobasal syndrome (CBS) is a rare Parkinsonian disorder that belongs to the group of tauopathies, in which neuroinflammation contributes to its pathophysiology.^1–3 CBS is characterized by marked asymmetry, rigidity, dystonia and apraxia, as well as cortical sensory and cognitive deficits.^2,4 Sleep quality and architecture (ie, the fundamental structure of normal sleep and the cyclical pattern of its various stages) are also affected in CBS; however, these aspects have not been extensively investigated. Small sample studies and case series have shown that sleep disorders, such as insomnia, restless leg syndrome, periodic limb movements, and sleep-wake cycle disturbances are common in CBS.^5–10 Additionally, alterations in sleep quality and architecture in these patients include reductions in total sleep time, sleep efficiency, non-rapid eye movement (NREM) and rapid eye movement (REM) sleep.^5,7,10 To our knowledge, no studies have quantified alterations in micro-sleep architecture (ie, transient, dynamic events and rhythms in neural electrical activity that occur within and across sleep stages) in CBS or examined whether sleep alterations may be linked to cognitive and motor decline, as seen in other neurodegenerative disorders.^11,12 Finally, little evidence exists regarding blood markers of systemic inflammation in CBS, despite the important contribution of (neuro) inflammation to τ (Tau) pathophysiology. More importantly, consistent evidence indicates that both regular exercise¹³ and sleep quality¹⁴ can positively influence systemic inflammation in different (clinical) populations.

Based on results from case reports in CBS^15–18 and other tauopathies,¹⁹ exercise has been recommended as an important component to counter gait and balance disturbances as well as cognitive deterioration. Moreover, exercise can enhance sleep quality and architecture in conditions with pathophysiological similarities to CBS, such as Parkinson disease (PD) and Alzheimer disease (AD), with multimodal training approaches potentially inducing the largest benefits.^20–22 Therefore, evidence that supports previously reported cross-sectional findings and longitudinal changes following regular exercise in this rare condition has important clinical implications. Here, we present a case report describing baseline and longitudinal changes in sleep quality, sleep architecture, cognition, motor function, fatigue and markers of systemic inflammation in an individual with CBS following 12 weeks of multimodal training.

METHODS

Research ethic approval was granted by the Comité d’Éthique de la Recherche (CÉR) CISSS de Laval (Project number: MP-50-2022-1584). This case report is reported following the CARE guidelines.²³

Case Description and Examination

One male individual, who was initially diagnosed with PD and subsequently re-diagnosed with CBS, gave informed consent to participate in the study (Table 1). The final diagnosis (ie, CBS) was obtained during the study period (ie, week 5) according to standard criteria and a positron emission tomography scan.^24–26 The participant was evaluated on anti-parkinsonian medications at the same time of the day (~1 hour) at baseline (T0) and after 12 weeks of multimodal training (T1). The evaluations were conducted over 2 days, separated by ~7 days at both time points. Blinded assessors conducted motor and cognitive evaluations (see below). The participant did not change medications during the study period.

Table 1.

Characteristics of the Participant at Baseline and Post-Training Changes^a

Demographics and Assessments	Baseline (T0)	Post-Intervention (T1)	Changes
Age	50.3 years	NA
Weight (BMI)	72.3 kg (23.1 kg/m2)	68.2 kg (21.8 kg/m2)	−4.1 kg
Current employment status	Not working	Not working	No change
Disease duration b	0.5 year	NA
Levodopa equivalent daily doses	600 mg	600 mg	No change
Pro oxazepam	15 mg	15 mg	No change
MDS-UPDRS III	31	42	+11 (deterioration)
PIGD sub-score	0	2	+2 (deterioration)
Tremor sub-score	4	9	+5 (deterioration)
TUG	12.7 s	11.0 s	−1.7 (improvement)
SCOPA-COG	25	34	+9 (improvement)
Memory and learning	11	16	+5 (improvement)
Attention	4	4	No change
Executive functions	0	9	+9 (improvement)
Visuospatial functions	10	5	−5 (deterioration)
RBDSQ-1 Item	0	0	No change
PDSS-2	10	15	+5 (deterioration)
BDI-2	11	14	+3 (deterioration)
PD Fatigue Scale	2.38	3.00	+0.62 (deterioration)
CPET
Test duration	12:00 min	14:00 min	+2:00 min
V̇O2peak	30.9 ml*kg−1*min−1	31.7 ml*kg−1*min−1	+0.8 ml*kg−1*min−1
Peak power output (W/kg)	155 W (2.14 W/kg)	165 W (2.42 W/kg)	+10 W (+0.28 W/kg)
Heart rate max (bpm)	159 bpm	163 bpm	+4 bpm
Rate of perceive exertion	7/10	9/10
Sub-maximal 1-RM test
Leg extension	58.4 kg	100 kg	+41.6 kg
Chest press	76.4 kg	90.7 kg	+14.3 kg
Lat machine	94.1 kg	100 kg	+ 5.9 kg
Leg curl	67.1 kg	NA
Inflammatory biomarkers
IL-1β	0.52 pg/ml	0.51 pg/ml	−0.01 pg/ml
IL-6	NAc	2.13 pg/ml	NA
IL-10	2.07 pg/ml	7.12 pg/ml	+5.05 pg/ml
TNFα	13.46 pg/ml	12.58 pg/ml	−0.88 pg/ml
CRP	4972.84 ng/ml	4761.86 ng/ml	−210.98 ng/ml

^a Bold text = main clinical or demographic variables. Abbreviations: BDI-2 = Beck Depression Inventory – 2; BMI = body mass index; bpm = beats per minute; CPET = cardiopulmonary exercise testing; CRP = C-Reactive Protein; IL-1β = Interleukin-1 β; IL-6 = Interleukin-6; IL-10 = Interleukin-10; MDS-UPDRS III = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale Part III; NA = not applicable; PD = Parkinson Disease; PDSS-2 = Parkinson’s Disease Sleep Scale version 2; PIGD = postural instability and gait difficulty; RBDSQ-1 = The REM sleep behavior disorder screening questionnaire – 1 item; RM = repetition maximum; s = seconds; SCOPA-COG = Scales for Outcomes in Parkinson’s Disease-Cognition; TNFα = Tumor Necrosis Factor α; TUG = Timed Up and Go test.

^b Disease duration of PD at the time of study enrollment.

^c Value lower than standard curve.

Therapeutic Intervention

The multimodal training combined resistance and cardiovascular training and was conducted at moderate to vigorous intensities for ~45 minutes, 3 times per week, over a 12-week period. Resistance training targeted large muscle groups (ie, chest, back, quadriceps, and hamstrings) and was performed with resistance machines. Cardiovascular training was performed on a total body recumbent stepper (ie, NuStep T4r, NuStep LLC, Plymouth, Michigan, USA), involving the upper and lower extremities. Each training session included 5 minutes of warm-up, followed by resistance training, cardiovascular training, and 5 minutes of cool-down. Training intensity and progression were tailored to the participant’s fitness level and determined using the percentage of heart rate (HR) max and the rate of perceived exertion (RPE) methods.²⁷ Training started at light/moderate intensity (50%–60% HRmax; 50%–55% 1-repetition max; RPE > 11) and progressively reached vigorous intensity (>80% HRmax; >80% 1-repetition max; RPE >14).²⁷ A detailed description of the intervention is reported in the Supplementary Material 1.

Outcome Measures

Functional Fitness Measures

Cardiorespiratory fitness level (V̇O_2peak) and muscle strength were assessed with a symptom-limited cardiopulmonary exercise test (CPET) using a metabolic cart (Quark CPET, Cosmed Srl, Italy) on a recumbent stepper and a submaximal 1-repetition maximum test, respectively. A detailed description of the functional fitness tests is reported in the Supplementary Material 2.

Sleep Quality and Architecture

Sleep quality was assessed subjectively with the PD Sleep Scale-2 (PDSS-2) and objectively through actigraphy (Actiwatch Spectrum Plus; Philips Respironics, Bend, OR, USA) worn on the non-dominant wrist for ~7 days, during which the participant also completed a sleep diary. Actiware software (v6.3; Philips Respironics, Bend, OR, USA) was implemented to automatically score the recording using an activity count threshold of 20, a minimum of 10 minutes of immobility, and rest durations of ≥30 minutes for major sleep bouts and ≥ 15 minutes for minor sleep bouts (up to 75 minutes).²⁸ Subsequently, all recordings were visually inspected, and the onset and offset of rest periods were cross-checked against the sleep diary. It should also be noted that actigraphy was used to monitor sleep patterns during the days preceding the overnight sleep evaluation conducted at our facilities. During this overnight evaluation, sleep architecture was assessed with polysomnography (Grael v2, Compumedics Limited, Abbotsford, Victoria, Australia). Specifically, electroencephalography (EEG) was conducted using a standard montage (10–20 international system) with frontal (F3-M2, F4-M1), central (C3-M2, C4-M1), parietal (P3-M2, P4-M1), and occipital (O1-M2, O2-M1) electrodes referenced to the mastoids. Electro-oculography and electromyography for chin muscle tone and tibialis anterior muscle were also collected. Signals were digitized at a sampling rate of 256 Hz (Profusion Sleep version 5, Compumedics Limited, Abbotsford, Victoria, Australia). An expert sleep technician, blinded to the study, manually scored the sleep recordings following American Academy of Sleep Medicine (AASM) guidelines using 30-second epochs.²⁹ Artifacts were detected by an automatic algorithm³⁰ and visually verified. In addition to macro-sleep architecture parameters, micro-sleep architecture (ie, sleep slow wave and spindle features) was examined during artifact-free NREM sleep stages 2 to 3 using Snooz Toolbox.³⁰ The same toolbox was implemented to conduct EEG power spectral analysis of low-frequency slow waves (≤4 Hz), slow oscillation (<1 Hz) and δ waves (1–4 Hz) throughout the sleep cycles. Sleep outcomes and methodologies implemented to analyze the EEG signal can be found in Supplementary Materials 3 and 4.

Cognition, Motor Function, Functional Mobility, and Fatigue

Cognition (ie, memory/learning, attention, executive function and visuospatial function domains) was assessed with the Scale for Outcomes in PD-Cognition (SCOPA-COG), with higher values reflecting better cognitive performance.³¹ Motor signs severity was assessed with the Movement Disorders Society Unified PD Rating Scale part III (MDS-UPDRS III), with higher scores indicating higher motor impairments.³² We also calculated the sub-scores reflecting postural instability/gait difficulty (ie, gait, freezing of gait, and postural stability; PIGD) and tremor.³³ Functional mobility was assessed using the Timed Up and Go (TUG) test.^34,35 The participant was instructed to sit on a chair with their back against the chair back. Upon the verbal cue “Go,” the participant was asked to stand up from the chair without using the armrests and with a comfortable, self-selected pace to walk to a cone placed 3 meters away, turn around it, return to the chair, and sit down. The TUG test was first demonstrated by a research assistant and was conducted after the participant practiced it once.^34,35 Fatigue was assessed with the PD Fatigue Scale, with higher scores indicating more severe fatigue.^36,37

Systemic Inflammation

The concentrations of interleukin (IL)-1β, IL-6, IL-10, tumor necrosis factor (TNF)α, and C-reactive protein (CRP) were measured from serum collected in the morning after a 12-hour fasting period. Blood samples were collected using a Vacutainer serum separator tube, allowed to clot for 1 hour at room temperature, cooled at ~4°C for 30 minutes, and centrifuged at 2200 g for 15 minutes. The resulting sera were aliquoted into 250 μL cryovials, stored at −80°C (Thermo Fisher Scientific, Waltham, MA, USA), and analyzed using Millipore Sigma kits (HSCTMAG-28SK and HCVD3MAG-67 K; Millipore Sigma, Saint Louis, Missouri, USA).

Role of the Funding Source

The funders played no role in the design, conduct, or reporting of this study. The work was supported by a Canadian Institutes of Health Research (CIHR) Project Grant (02109PJT468982-MOV-CFAA-244681). J.C.C. received funding from the Brain Canada Next Gen Award in Parkinson’s Disease Research, and the Fonds de recherche du Quebec (FRQS) (doctoral scholarships). M.R. was supported FRQS Salary Award (Junior 2).

RESULTS

The participant completed all the evaluations and 34 training sessions (total training time: 25 h 21 min; average rate of perceived exertion during training = 16/20) without experiencing any adverse events. Tables 1, 2, and 3 show baseline and post-training outcomes. After the exercise program, cardiorespiratory fitness level (+0.8 ml^*kg^−1*min⁻¹) and sub-maximal strength (Table 1) improved. The participant reported a subjective reduction in sleep quality (PDSS-2: 10/60 to 15/60) post intervention. Despite this reduction in subjective sleep quality, actigraphy and polysomnography showed improvements in several sleep outcomes after the exercise program (Table 2). That is, actigraphy results showed an improvement in total sleep time (383.4 [65.0] to 433.3 [82.3] min), sleep efficiency (81.8 [9.7]% to 90.6 [3.9]%) and sleep onset latency (16.6 [16.4] min to 2.5 [4.1] min). Additionally, there was a reduction in daily activity counts (222.9 [38.8] AC/min to 107.5 [60.1] AC/min), suggesting a reduction in daily physical activity levels post-treatment.

Table 2.

Actigraphy and Macro-Sleep Architecture Outcomes at Baseline and After 12 Weeks of Training^a

Actigraphy	Baseline (T0)	Post-Intervention (T1)
Nights recorded	9	6
TST, min, avg (SD)	383.4 (65.0)	433.3 (82.3)
WASO, min, avg (SD)	31.3 (13.1)	29.4 (14.7)
SOL, min, avg (SD)	16.6 (16.4)	2.5 (4.1)
Sleep efficiency, %, avg (SD)	81.8 (9.7)	90.6 (3.9)
Sleep fragmentation, %, avg (SD)	5.8 (5.7)	6.5 (3.2)
Daily Activity Count, AC/min, avg (SD)	222.9 (38.8)	107.5 (60.1)
Polysomnography: Sleep Architecture	Baseline	Post-Intervention
TIB, min	271	487.5
TST, min	193	461
Sleep efficiency, %	71.2	94.6
SOL, min	33	9
REM latency, min	167	158
WASO, min	78	26.5
N1, min	8.5	17
N2, min	162	334.5
N3, min	3.5	4.5
NREM, min	174	356
REM, min	19	105
Wake, %	28.8	5.4
N1, %	4.4	3.7
N2, %	83.9	72.6
N3, %	1.8	1.0
NREM, %	90.2	77.2
REM, %	9.8	22.8

^a Bold text = sleep assessments. Abbreviations: avg = average; AC = activity count; N1 = NREM stage N1 sleep; N2 = NREM stage N2 sleep; N3 = NREM stage N3 sleep; NREM = non-rapid eye movement; REM = rapid eye movement; SOL = sleep onset latency; SD = standard deviation; TIB = time in bed; TST = total sleep time; WASO = wake after sleep onset.

Table 3.

Slow Waves and Spindle Features at Baseline and After 12 Weeks of Training^a

Slow Waves and Spindle Features	Baseline (T0)	Post-Intervention (T1)
Slow wave (count)	Frontal: 1846 Central: 1008 Parietal: 373 Total: 3234	Frontal: 1407 Central: 628 Parietal: 215 Total: 2259
Slow wave density (count/min)	Frontal: 5.6 Central: 3.0 Parietal: 1.1 Total: 2.8	Frontal: 2.1 Central: 0.9 Parietal: 0.3 Total: 1.0
Slow wave amplitude (μV)	Frontal: 102.3 Central: 98.8 Parietal: 93.4 Total: 96.3	Frontal: 100.2 Central: 96.3 Parietal: 91.1 Total: 94.5
Slow wave slope (min-to-max; μV/sec)	Frontal: 358.2 Central: 356.8 Parietal: 306.8 Total: 319.9	Frontal: 271.6 Central: 246.8 Parietal: 202.6 Total: 224.5
Slow wave frequency (Hz)	Frontal: 1.4 Central: 1.5 Parietal: 1.3 Total: 1.3	Frontal: 1.1 Central: 1.0 Parietal: 0.9 Total: 1.0
Slow wave transition frequency (Hz)	Frontal: 1.7 Central: 1.8 Parietal: 1.6 Total: 1.6	Frontal: 1.4 Central: 1.3 Parietal: 1.1 Total: 1.2
Spindle (count)	Frontal: 513 Central: 171 Parietal: 77 Total: 770	Frontal: 2720 Central: 1777 Parietal: 1279 Total: 5927
Spindle density (count/min)	Frontal: 1.6 Central: 0.5 Parietal: 0.2 Total: 0.6	Frontal: 4.0 Central: 2.6 Parietal: 1.9 Total: 2.2
Spindle amplitude (μV)	Frontal: 29.9 Central: 25.9 Parietal: 24.0 Total: 25.9	Frontal: 30.2 Central: 25.4 Parietal: 23.0 Total: 24.7
Spindle frequency (Hz)	Frontal: 12.5 Central: 12.7 Parietal: 13.4 Total: 13.0	Frontal: 12.6 Central: 13.1 Parietal: 13.6 Total: 13.1

Consistent with the actigraphy results, polysomnography results showed an increase in total sleep time (195.0 min to 462.5 min), sleep efficiency (71.2% to 94.6%), N2 (162 min to 334.5 min), REM (19 min to 105 min) and a reduction in sleep onset latency (33.0 min to 9.0 min). Furthermore, while no major changes in N1% (4.4% to 3.7%) and N3% (1.8% to 1.0%) were found, there was a reduction in N2% (83.9% to 72.6%) and wake time (28.8% to 5.4%), as well as an increase in REM% (9.8% to 22.8%). It should be noted that sleep stage N3 was extremely short and did not change post-exercise (Table 2). Hypnograms can be found in Supplementary Material 5.

Sleep spindles and slow wave features (ie, micro-sleep architecture) are reported in Table 3, and power spectral analysis results are shown in the Figure 1. Sleep spindle density increased (average across channels: 0.58 count/min to 2.20 count/min) post-exercise. In contrast, slow wave density (average across channels: 2.8 count/min to 1.0 count/min) as well as slow wave slope, frequency, and transition frequency decreased. These changes in slow waves indicate potential (post-exercise) modifications not only in slow wave density but also in morphology. Finally, there was an overall reduction in low-frequency slow wave power across sleep cycles post-exercise (Figure 1A). Importantly, however, this overall reduction was specific to δ wave power (Figure 1B) and did not occur for slow oscillations, whose power increased after exercise (Figure 1C). A similar power reduction was observed for fast δ waves power (2–4 Hz³⁸; Suppl. Material 6).

Figure 1.

Power Spectral Analysis. (A) Low-Frequency Slow Wave Power—T0: Cycle 1 = 259.6 μV²; Cycle 2 = 174.6 μV.² T1: Cycle 1 = 202.4 μV²; Cycle 2 = 116.5 μV²; Cycle 3 = 67.8 μV²; Cycle 4 = 75.6 μV.² (B) Slow Oscillation Power—T0: Cycle 1 = 51.8 μV²; Cycle 2 = 39.3 μV.² T1: Cycle 1 = 70.4 μV²; Cycle 2 = 49.2 μV²; Cycle 3 = 20.2 μV²; Cycle 4 = 19.3 μV.² (C) δ Power—T0: Cycle 1 = 207.8 μV²; Cycle 2 = 135.3 μV.² T1: Cycle 1 = 132.0 μV²; Cycle 2 = 67.3 μV²; Cycle 3 = 47.6 μV²; Cycle 4 = 56.3 μV.² Abbreviations: EEG = Electroencephalography; NREM = Non-rapid Eye Movement Sleep; T0 = Baseline; T1 = Post-Intervention.

Following the training program, an overall improvement in cognition (SCOPA-COG: 25/43 to 34/43) and functional mobility (TUG: 12.7 seconds to 11.0 seconds) were observed. However, the cognitive improvement was specific to memory/learning and executive functions domains while a worsening in visuospatial functions was detected. Additionally, the improvement in TUG did not exceed minimal detectable changes and was lower than the standard errors of measurement estimated for individuals with PD and AD.^35,39 Conversely, the participant experienced a deterioration in motor signs (MDS-UPDRS III: 31/132 to 42/132) and fatigue (PD Fatigue Scale: 2.38/5 to 3/5). Finally, peripheral TNFα and CRP blood concentrations decreased, while IL-10 increased post intervention (Table 1).

DISCUSSION

The participant reported relatively good satisfaction with their subjective sleep perception at baseline. However, there were marked alterations in objective sleep quality and architecture, confirming previously observed sleep disturbances in CBS.^5,7,9,10 That is, our participant showed reduced sleep time and efficiency, NREM and REM sleep, and nearly absent N3 (ie, slow wave sleep). Furthermore, we quantified important sleep slow wave and spindle density deteriorations, with values markedly lower than those reported in adults without neurological disorders^40,41 and, in case of sleep spindles, even lower than those reported in PD.^42,43 However, the deterioration of these neural oscillations was broadly comparable to that observed in other tauopathies, such as AD.¹² Notably, following the 12-week multimodal training program, we observed improvements in multiple sleep outcomes measured objectively both with actigraphy over several days in the participant’s home setting as well as with polysomnography. However, we cannot rule out the possibility that a first-night effect could have biased the polysomnography outcomes. These results suggest that our participant may have experienced an improvement in objective sleep quality, particularly in sleep efficiency,⁴⁴ following the training program. It is important to emphasize, however, that our observations should be interpreted with caution, and do not provide evidence supporting a causal effect of exercise in CBS.

Conversely, we observed a deterioration in self-reported subjective sleep quality post-exercise intervention. This deterioration is well above the minimal clinically important difference (MCID; 2.07 points) established for PD⁴⁵ and suggests a discrepancy between subjective and objective sleep measures. The reasons for the observed decline in subjective sleep quality remain unclear. Potential explanations include sleep misperception (already evident at baseline; PDSS-2 = 10/60), increased awareness/expectations following the intervention, a slight worsening of fatigue and depressive symptoms (Table 1), or a combination of these factors.^46–50 However, this discrepancy is not novel, it has been observed repeatedly in other (clinical) populations, and it underlies the importance of combining objective and subjective sleep measures to capture all the dimensions of sleep.^20,21,51 Additionally, it reinforces the notion that a combination of different interventions (eg, exercise and sleep hygiene) might be required to improve all the dimensions of sleep.^20,21,48

In addition to improvements in objective sleep quality, we observed changes in features of micro-sleep architecture and low-frequency slow wave power. Specifically, sleep spindle density, which was particularly low at baseline,⁴¹ markedly increased following the training program, corroborating similar positive effects observed in PD after a comparable training intervention.⁵² Notably, overall post-training sleep spindle density reached values typically observed in adults without neurological disorders (ie, ~2–3/min).⁴¹ It would be important to determine whether the observed increase in sleep spindle density is solely the results of longer sleep duration and time spent in sleep stage N2 during the second evaluation or linked to other mechanisms.^41,53

Conversely, we observed a reduction in slow wave count/density following the training program. It is also important to note that slow wave density was markedly low at baseline compared to previously published data from adults without neurological disorders in their middle age⁴⁰ but similar to what reported in PD⁴³. These results expand prior research in CBS, suggesting an early disruption of slow waves and spindles presumably induced by neurodegeneration within sleep-regulatory brain structures observed in tauopathy.^5,7,10,54 Whether regular exercise can improve some of these relevant neural oscillations (ie, sleep spindles and slow waves) in CBS remains unclear at this point, and future controlled studies are required.

The reduction in slow wave (count/density) was paralleled by a reduction in low-frequency slow waves power, particularly during the first 2 sleep cycles after the training program (Figure 1A). However, this reduction appeared to be specific to the δ power (Figure 1C) while slow oscillations power increased (Figure 1B), a finding that is consistent with the changes also observed in slow wave morphology (frequency/transition frequency; Table 3). Slow oscillations and δ waves are thought to be regulated and respond differently to homeostatic sleep pressure (ie, the need for sleep accumulates with prolonged wakefulness and dissipates during sleep).^38,55–58 Given that during the days leading up to the second polysomnography evaluation, the participant experienced a substantial improvement in objective sleep quality and a potential overall reduction in levels of physical activity measured with actigraphy (Table 1), 1 could speculate that an overall reduction in homeostatic sleep pressure might be associated with the reduction in (fast) δ power. In any event, the effects of regular exercise on those power frequencies and their clinical relevance in CBS warrant further investigation in controlled studies.

At baseline, the participant scored 25 on the SCOPA-COG, which is slightly above the established cut-off (<25) for identifying mild cognitive impairment in PD,⁵⁹ indicating some cognitive dysfunction. Executive and visuospatial dysfunctions are early, key cognitive alterations that can be implemented as criteria for diagnosing CBS.²⁴ Our longitudinal observation suggests that regular exercise, as reported in other neurodegenerative disorders,^60,61 may be associated with improvements in overall cognitive function in CBS.¹⁹ However, the absence of a control case precludes any causal interpretation. Notably, a closer analysis indicates that the overall cognitive enhancement was domain-specific (ie, executive function and memory/learning domains), while some aspects of cognition (ie, visuospatial functions) kept deteriorating. Controlled studies are required to determine whether regular exercise improves specific cognitive domains in CBS and whether changes in sleep architecture may mediate this cognitive improvement.^{11,12,43,52,62–66}

We also observed increased fatigue, reduction in physical activity levels, and a marked deterioration in motor signs, whereas functional mobility (ie, TUG) remained stable following the training program. Fatigue is a detrimental feature common in Parkinsonian syndromes,^67–69 and the perceived increase in fatigue could be related to the observed deterioration in other non-motor symptoms (ie, subjective sleep quality and depression).^49,50 Furthermore, our observation partially aligns with previous RCTs in PD, showing that regular exercise is not always effective at improving feelings of fatigue.^67,68 Several non-mutually exclusive factors might explain the observed reduction in physical activity levels following the training program. While this reduction may not to be associated with a deterioration in mobility (ie, TUG), it may be linked to seasonal temperature changes,⁷⁰ as the baseline and post-exercise evaluations took place in July (average 23.6°C) and November (average 4.9°C), respectively. Additionally, it may be associated with the observed increased levels of fatigue or the training program itself.^71–73 That is, prior exercise studies in adults without neurological disorders and those with PD have shown that daily engagement in non-exercise activities, such as sedentary behaviours, can increase as a compensatory response to progressive, intensive training programs.^71–75 Given that inactivity is an independent risk factor for all-cause mortality,⁷⁶ further research is warranted, and complementary behavioral strategies may be necessary to help sustain adequate physical activity levels during an exercise program.⁷⁷

The observed motor decline (MDS-UPDRS III) following the training program exceeded the MCID (4.63 points) to detect worsening in PD.⁷⁸ This observation contrasts with the positive effects of regular exercise on motor function in PD^79,80 and the preserved functional mobility (ie, TUG) observed in our participant following the training program. Several concomitant factors could partially explain this deterioration. These may include the potential progression of the disease (ie, PIGD and Tremor sub-scores),⁸¹ poor response to levodopa in CBS, well-established limitations of the MDS-UPDRS III, and the type of training modality used.^{2,4,19,81–84} It would also be crucial to determine whether this motor deterioration resulted from increased fatigue, which our intervention might have triggered. This hypothesis, however, contrasts with (meta-analytic) findings in PD and progressive supranuclear palsy,^49,50,85,86 which suggest that fatigue is an early, consistent, and independent non-motor symptom not necessarily associated with motor progression. Moreover, this interpretation appears inconsistent with the observed fitness improvements following the training program⁸⁶ and the fact that the intervention was both tailored and progressive. Another important contributor to the motor deterioration might be linked to the abnormal sleep architecture observed in our participant. As δ wave activity (ie, power) can predict motor progression in PD,^11,87 it is crucial to investigate whether similar associations can be observed in other neurodegenerative disorders, including CBS, over a more extended period.

Unlike the motor decline observed in MDS-UPDRS III scores, functional mobility, as measured by the TUG, remained stable following the training program. Given the absence of a control and the potential for practice effects, it is not possible to determine whether our observations hold clinical significance. However, they align with previous findings in CBS and PD,^19,88 suggesting that regular exercise may be beneficial for functional mobility in Parkinsonian syndromes. Notably, this finding contrasts with the changes in the PIGD sub-score, suggesting that more objective measures of motor function and functional mobility may be needed to support or complement changes detected by the MDS-UPDRS III over short periods.⁸¹

Finally, the systemic inflammatory profile of our participant may have improved following the training program. Although our case report precludes any causal inference, the observed reduction in systemic inflammation may be associated with exercise,¹³ improved sleep quality,¹⁴ or both mechanisms. Notably, previous evidence suggests that exercise may enhance sleep quality and architecture in neurological disorders such as PD and AD by reducing inflammation.²² Thus, future studies, including a non-exercise control condition or multiple baseline evaluations, should determine whether exercise improves systemic inflammation in CBS and whether this change leads to or results from better sleep.

Limitations

We acknowledge that this case report has limitations. First, although our observations align with previous findings in this clinical population,¹⁹ they cannot be generalized to all individuals with CBS. Second, the lack of a control condition limits the strength of our observations and conclusions. We interpreted our results considering MCID and established cut-off values to partially address this limitation. Finally, the absence of follow-up prevents us from assessing whether and for how long the observed positive changes were sustained.

In summary, this case report supports the notion that objective sleep quality, sleep architecture, and cognitive function are markedly altered in CBS. It further suggests that moderate-to-vigorous exercise is feasible and may improve fitness, objective sleep quality and architecture, overall cognition, gait performance, and systemic inflammation in an individual with CBS. However, subjective sleep quality, sleep slow waves, motor signs and fatigue worsened following the training program. While these findings partially support previous reports in CBS and the notion that regular exercise can help manage Parkinsonian syndromes, controlled studies are warranted to confirm our observations and interpretations.

CRediT—CONTRIBUTOR ROLES

Jacopo Cristini (Conceptualization [Equal], Data curation [Lead], Formal Analysis [Lead], Methodology [Lead], Project administration [Lead], Supervision [Equal], Writing—original draft [Lead], Writing—review & editing [Lead]), Freddie Seo (Investigation [Equal], Project administration [Equal], Writing—review & editing [Supporting]), Ashanté Bon (Investigation [Equal], Project administration [Equal], Writing—review & editing [Supporting]), Lynden Rodrigues (Investigation [Equal], Project administration [Equal], Writing—review & editing [Equal]), Kira Sikorska (Data curation [Equal], Investigation (Equal), Writing—review & editing [Equal]), Anke Van Roy (Writing—review & editing [Equal]), Alain Dagher (Writing—review & editing [Equal]), Julie Carrier (Writing—review & editing [Equal]), Caroline Paquette (Investigation [Equal], Project administration [Equal], Resources [Equal], Supervision [Equal], Writing—review & editing [Equal]), and Marc Roig (Conceptualization [Equal], Funding acquisition [Lead], Project administration [Equal],Resources [Equal], Software [Equal], Supervision [Equal], Writing—review & editing [Equal]).

FUNDING

The work was supported by a Canadian Institutes of Health Research (CIHR) Project Grant (02109PJT468982-MOV-CFAA-244681). J.C.C. received funding from the Brain Canada Next Gen Award in Parkinson’s Disease Research, and the FRQS (doctoral scholarships). M.R. was supported FRQS Salary Award (Junior 2).

ETHICS APPROVAL

Research ethic approval was granted by the Comité d’Éthique de la Recherche CISSS de Laval (Project number: MP-50-2022-1584).

DISCLOSURES

The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.

A preprint of an earlier version of this manuscript is available on Research Square (DOI: http://dx.doi.org/10.21203/rs.3.rs-6696024/v1).

DATA AVAILABILITY

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.