Higher Striatal Dopamine is Related With Lower Physical Performance Fatigability in Community-Dwelling Older Adults

Abstract

Background

Fatigability in community-dwelling older adults is highly prevalent and disabling, but lacks a treatment. Greater nigrostriatal dopaminergic signaling can ameliorate performance fatigability in healthy young adults, but its role in community-dwelling older adults is not known. We hypothesized that higher nigrostriatal dopaminergic integrity would be associated with lower performance fatigability, independent of cardiopulmonary and musculoskeletal energetics and other health conditions.

Methods

In 125 older adults participating in the Study of Muscle, Mobility and Aging, performance fatigability was measured as performance deterioration during a fast 400 m walk (% slowing down from the 2nd to the 9th lap). Nigrostriatal DA integrity was measured using (+)-[11C] dihydrotetrabenazine (DTBZ) PET imaging. The binding signal was obtained separately for the subregions regulating sensorimotor (posterior putamen), reward (ventral striatum), and executive control processes (dorsal striatum). Multivariable linear regression models of performance fatigability (dependent variable) estimated the coefficients of dopamine integrity in striatal subregions, adjusted for demographics, comorbidities, and cognition. Models were further adjusted for skeletal muscle energetics (via biopsy) and cardiopulmonary fitness (via cardiopulmonary exercise testing).

Results

Higher [¹¹C]-DTBZ binding in the posterior putamen was significantly associated with lower performance fatigability (demographic-adjusted standardized β = −1.08, 95% CI: −1.96, −0.20); results remained independent of adjustment for other covariates, including cardiopulmonary and musculoskeletal energetics. Associations with other striatal subregions were not significant.

Discussion

Dopaminergic integrity in the sensorimotor striatum may influence performance fatigability in older adults without clinically overt diseases, independent of other aging systems.

Greater fatigability, the degree of performance deterioration during a standardized task, is common and disabling in community-dwelling older adults, and predicts mortality (1–3). While studies in older adults show associations of fatigability with lower muscle mitochondrial function and cardiopulmonary fitness, these factors only partially explain fatigability (4,5). Strategies to ameliorate muscle mitochondrial function and cardiopulmonary fitness have shown beneficial effects on fatigability in patient populations with specific disorders; however, there is currently no treatment when fatigue arises without a specific cause or disease.

There is evidence for a role of dopaminergic neurotransmission in endurance and fatigability, albeit limited to athletes and healthy young adults. In healthy young adults, greater dopaminergic signaling promotes performance by dampening the perception of the physiological markers of fatigue (rising heart and respiratory rate, muscle cramps) while increasing motivation and attention (6,7). Emerging evidence suggests dopaminergic neurotransmission is positively related with physical performance (8) and with lower perceived fatigability in older adults without neurological diseases (9–11). However, studies primarily used genetic polymorphism or neuroimaging markers of striatal gray matter volume and connectivity, which are indirect markers of dopaminergic signaling, and did not account for other physiological contributors of fatigability. While it is biologically plausible that dopaminergic neurotransmission would influence performance fatigability in older adults, this hypothesis has not been directly tested. It is also not known whether the influence of dopaminergic signaling on performance fatigability would be independent of other health conditions, and in particular muscle energetics and cardiopulmonary fitness.

We examined striatal dopaminergic integrity using a well-established radioligand in older adults without neurological diseases (12) and with validated measures of performance fatigability, as well as comprehensive imaging and biopsy-based data on skeletal muscle energetics, and cardiopulmonary fitness. We hypothesized that higher nigrostriatal dopaminergic integrity would be associated with lower performance fatigability independent of muscle and cardiopulmonary contributors.

Method

Study Sample

A total of 150 community-dwelling adults aged ≥70 years were recruited from the Study of Muscle, Mobility and Aging (SOMMA, https://sommaonline.ucsf.edu) at the University of Pittsburgh for the Brain Ancillary study. Exclusion criteria for the parent study were: inability to walk one-quarter of a mile or climb a flight of stairs; had BMI ≥ 40 kg/m²; had an active malignancy or dementia; or any medical contraindication to biopsy or MRI. Finally, participants must have been able to complete the 400 m walk at usual pace; those who appeared as they might not be able to complete the 400 m walk at the in-person screening visit completed a short distance walk (4 m) to ensure their walking speed was ≥0.6 m/s (13). Participant enrollment and baseline testing started in April 2019 and were completed in December 2021. The SOMMA Brain ancillary study included SOMMA participants at the Pittsburgh site who received a muscle biopsy in the prior 12 months. Among 439 SOMMA participants at the Pittsburgh site, 285 had a muscle biopsy within the prior 12 months; of these, 150 underwent neuroimaging, 2 were ineligible and the others were not interested. After neuroimaging, 125 returned for the fast-paced 400 m walk and other exams, the others could not be reached (n = 2), refused (n = 1), began but did not complete the walk (n = 2), were ineligible (n = 1), or the test was not offered due to scheduling issues (n = 19). The last lap of one participant was accidentally deleted; therefore, the final analytical sample size was 124. The interval of time between the muscle biopsy and neuroimaging on average was 9.4 months. The WIRB-Copernicus Group Institutional Review Board approved the study and all participants gave informed written consent (parent study and ancillary numbers 20180764 and 20110230).

Performance Fatigability

Study participants walked a 400-m course in a dedicated long corridor, with traffic cones on both ends. Participants were instructed to walk “as quickly as possible, without running, at a pace they could maintain” for 10 laps (20 m each direction). Start and stop time of the walking task was collected as well as time to completion (in seconds) and split times for each of the 10 laps (in seconds) (14). Standardized encouragement was provided at the beginning of each lap as follows: “You are doing a good job. You have completed XX laps and have XX laps to go.” A test was stopped if a participant reported chest pain, tightness, pressure in the chest, shortness of breath, leg pain; or feeling faint, lightheaded, dizzy; or they requested to stop (14).

Percent change in gait speed during the fast 400-m walk was calculated as the inverse value of [(lap 2 − lap 9)/lap 2] × 100 so that a higher value reflects greater performance deterioration (ie, slowing down);14 a measurement of performance fatigability with good validity (15). In addition to using performance fatigability as a continuous measure, we also calculated a binary variable using a cutoff of ≥6.5% to identify those with marked performance deterioration (ie, severe performance fatigability), as previously described in a similarly aged cohort (14).

Structural Brain Imaging

Structural magnetic resonance (MR) images were collected on a Siemens Biograph mMR PET/MR system capable of simultaneous PET and 3 Tesla (3T) MR image acquisition. 3T MR images were acquired simultaneously with PET image acquisition and included T1-weighted magnetization-prepared rapid gradient echo (MPRAGE) and Dixon sequences (16). Distortion correction, registration, and segmentation were obtained via a combination of FreeSurfer and FSL (17) utilizing the minimum preprocessing pipeline (18) and the Human Connectome Project workbench software (version 1.4.0) (19). White matter hyperintensities were mapped using the T2-FLAIR MRI volume via an automated segmentation method (20) that uses fuzzy connectedness to identify hyperintensities. The main measures of interest for these analyses were gray matter volume from total brain, normalized by intracranial volume, and white matter hyperintensity volumes, normalized by total brain volume.

PET Imaging Protocol

We examined the integrity of nigrostriatal dopaminergic terminals using a well-established PET radioligand targeting the type-2 vesicular monoamine transporter (VMAT2), [¹¹C]-α-(+)-dihydrotetrabenazine ([¹¹C]DTBZ) (12). VMAT2 is expressed in presynaptic vesicular membranes and transports monoamine neurotransmitters, including dopamine, from the cytosol to storage vesicles. While not specific to dopamine, nearly all (>95%) striatal VMAT2 is associated with storage vesicles in dopamine terminals. Additionally, relative to other presynaptic dopamine targets (eg, dopamine transporter, aromatic amino-acid decarboxylase), VMAT2 is less affected by pharmacologic or lesion-compensatory regulation (21,22). In the context of (1) minimal regulation of VMAT2 by dopamine enhancing drugs or compensatory processes, and (2) absence of neurological diseases like Parkinson’s disease as is the case in this aging population, higher [¹¹C]DTBZ striatal binding in vivo neostriatum (caudate nucleus and putamen) and the limbic striatum (anterior ventral striatum) corresponds to higher levels of storage vesicles in the dopamine terminals originating from the subtantia nigra pars compacta and the ventral tegmental area, respectively, and it suggests greater nigrostriatal dopaminergic integrity and signaling capacity. Therefore, we refer to higher/lower [¹¹C] DTBZ binding in PET images to indicate higher/lower dopamine nerve terminal integrity.

For each participant, [¹¹C]DTBZ (10–15 mCi) was injected intravenously as a slow bolus (20–30 seconds) and list-mode acquisition of PET emission data commenced at the start of radiotracer injection and continued for 60 minutes as previously described (12). PET emission data were reconstructed using filtered back projection with Fourier rebinning into a dynamic series of 20 frames ranging in duration from 15 to 300 seconds. Attenuation correction of PET emission data was performed analytically using previously described model-based methods (23,24).

Image Analyses

PMOD software (v 3.7, Zurich, Switzerland) was used for PET data analyses, including motion correction, PET to MR image registration, segmentation, sampling, and tracer kinetic analyses. Volumes of interest for sampling of PET image data were defined on the MPRAGE series using the Imperial College London Clinical Imaging Center atlas, including the functional subdivisions of the striatum of interest (CIC atlas (25)): associative (the volume-weighted average of [¹¹C]DTBZ activity in the anterior putamen, pre-dorsal caudate, and post-dorsal caudate), limbic (anterior ventral striatum) and sensorimotor (posterior putamen). Quantitation of [¹¹C]DTBZ specific-binding outcomes (BP_ND) was determined for each volume of interest using the noninvasive Logan graphical analysis method and occipital cortex as a reference tissue (26–28).

Skeletal Muscle Respiration

The skeletal muscle biopsy was taken from the medial vastus lateralis (right side, or left if necessary) after a 12-hour fast and limited exercise for 48 hours prior to the procedure (13). A portion (~20 mg) of the specimen was placed in a biopsy-preserving solution for high-resolution respirometry, as previously described (29). Myofiber bundles (~2–3 mg) were prepared by gently separating the fibers using fine-point tweezers. After weighing, the permeabilized myofiber bundles were placed into Oxygraph-2K respirometer chambers (O2K, Oroboros Instruments, Austria). Assays were run at 37°C within a specific range of O₂ concentration (400–200 μm). Steady-state oxygen flux was normalized to the fiber bundle wet weight using Datlab 7.4 software. Technician was controlled for in analysis.

Cardiorespiratory Fitness

Cardiorespiratory fitness was measured by cardiopulmonary exercise testing using a modified symptom-limited Balke treadmill protocol where speed and grade increased incrementally (30,31). Ventilatory gases, oxygen and carbon dioxide were inhaled and exhaled during exercise. VO₂peak was standardized for weight for this analysis (mL/kg/min). The data were manually reviewed to ensure the correct VO₂peak was selected for each participant. Out of the 875 participants who completed the CPET, 20 (or 2.28%) had an invalid test and 855 participants (97.7%) had a valid test (30).

Other Variables of Interest

Information on age in years was recorded at the time of baseline MRI, the other variables were from the baseline visit of the parent SOMMA. The Montreal Cognitive Assessment (MoCA, range 0–30, higher = better) was obtained to assess attention, concentration, executive function, memory, language, visuo-constructional skill, conceptual thinking, calculations, and orientations. Multimorbidity was assessed via self-reported physician diagnosis of 11 conditions (cancer, cardiac arrhythmia, chronic kidney disease, chronic obstructive pulmonary disease, coronary heart disease, congestive heart failure, and dementia) and summed for analysis. A count of prescription medications included anti-dopaminergic, anticholinergic, and central nervous system active medications that could affect mobility. Perceived physical fatigability was assessed using the Pittsburgh Fatigability Scale (PFS, range 0–50, higher = greater fatigability) (32). Body mass index (BMI, kg/m²) was calculated from height (Harpenden stadiometer, Dyved, UK) and weight (balance beam or digital scale), measured without shoes and with light clothing. Joint pain was assessed via self-reported lower extremity pain in the last 12 months for 30 days, and dichotomized (yes/no) for analysis.

Statistical Analysis

Characteristics of the SOMMA Brain cohort were compared to the SOMMA parent cohort at the Pittsburgh site, using t tests for continuous variables and chi-squared tests for categorical variables. Pearson’s (continuous) or Spearman (categorical) correlations were generated to estimate the bivariate association between population characteristics and performance deterioration. Because skeletal muscle energetics (maxOXPHOS) and cardiorespiratory fitness (VO₂peak) are known contributors of fatigability, we also assessed the association of ¹¹[¹¹C]-DTBZ binding in striatal subregions with these variables. Linear regression models were generated to test the association between dopamine neurotransmission in the 3 striatal subregions (sensorimotor, associative, reward, and entered singly) and performance deterioration, adjusted for age and sex. Striatal region(s) with a statistically significant coefficient entered a model further adjusted for other covariates; models with muscle energetics were further adjusted for technician. A parsimonious model included dopamine neurotransmission, age, sex, and other variables that were bivariately associated with performance deterioration. A final model included variables with known associations with the outcome based on literature review: VO₂peak, BMI, joint pain, and medications. All variables were standardized (1 standard deviation increments), and all models included R² to appreciate the contribution of dopamine and other variables to performance deterioration. In sensitivity analyses, we adjusted for overall speed during the fast 400 m walk; additionally, we used logistic regression models with marked performance deterioration (ie, severe performance fatigability) as the outcome. Analyses were conducted in SAS version 9.4.

Results

Among the participants at the SOMMA-Pittsburgh site, those included in this analysis were younger, with lower levels of perceived physical and mental fatigability, lower BMI, and higher skeletal muscle energetics and cardiopulmonary fitness (Table 1).

Table 1.

Characteristics of Participants of the Study of Muscle, Mobility and Aging (SOMMA) at the Pittsburgh Field Center Site, Included in This Analysis and Those Excluded

Variable	Included (n = 124)	Excluded* (n = 311)	p Value
	Mean ± Standard Deviation, or N (%)
Performance deterioration, %	2.5 ± 4.7	NA
Time to walk fast 400 m, s	305.0 ± 59.8	NA
Age, years	75.4 ± 4.0	76.9 ± 5.1	.0025
Women	74 (59.7)	191 (61.4)	.7
maxOXPHOS, pmol/(s*mg)	67.7 ± 20.5	56.7 ± 17.4	<.001
VO2peak, mL/kg/min	23.1 ± 5.5	20.0 ± 4.9	<.001
Body mass index, kg/m2	26.5 ± 4.1	28.1 ± 4.8	.0014
Multimorbidity index > 1	18 (14.6)	52 (16.9)	.6
Pittsburgh Fatigability Scale Physical score, range 0–50	12.5 ± 7.4	15.8 ± 8.3	.0001
MoCA, range 0–30	25.2 ± 2.5	24.9 ± 2.7	.6
Joint pain, present	41 (33.1)	118 (37.9)	.3
Medications, number	4.6 ± 3.6	4.7 ± 3.8	.8

Notes: NA = not available as measures were only collected in SOMMA Brain ancillary; MoCA = Montreal Cognitive Assessment; maxOXPHOS = Maximal Complex I&II Supported OXPHOS; VO₂peak = peak oxygen consumption. Performance deterioration was calculated as the inverse value of [(lap 2 − lap 9)/lap 2] × 100; higher values reflect greater slowing down (ie, greater performance fatigability).

^* n = 311 calculated as total N = 439 enrolled in SOMMA Pittsburgh minus 128 (ie, 4 missing 400 m walk time and the 124 in SOMMA brain ancillary analytic sample).

Performance deterioration was on average 2.5%, ranging from a gait speed decline of 16% to an increase of 13% from the 2nd to the 9th lap. Marked fatigability (performance deterioration ≥6.5%) was identified in 20 of 124 participants (16.7%). Lower MoCA scores and higher maxOXPHOS were correlated with greater performance deterioration (Table 2). Associations with cardiopulmonary fitness and health-related indices (BMI, multimorbidities, medication use, etc.) were not statistically significant (Table 2). The associations of striatal ¹¹[11C]-DTBZ binding with maxOXPHOS, VO₂peak, depressive symptoms, or apathy were not significant (p > .5).

Table 2.

Correlations Between Variables of Interest and Performance Deterioration in 124 SOMMA Participants

Variable	R *	p Value
Time to walk fast 400 m, s	0.05	.550
Age, years	0.12	.172
Women	0.02	.835
maxOXPHOS, pmol/(s*mg)	0.20	.034
VO2peak, mL/kg/min	-0.06	.528
Body mass index, kg/m2	-0.06	.475
Multimorbidity index > 1	-0.02	.842
Pittsburgh Fatigability Scale Physical score, range 0–50	0.02	.858
MoCA, range 0–30	-0.23	.010
Joint pain, present	0.08	.352
Medications, number	0.09	.320
Gray matter volume†	0.05	.607
White matter hyperintensities‡	0.07	.466

Notes: MoCA = Montreal Cognitive Assessment; maxOXPHOS = Maximal Complex I&II Supported OXPHOS; SOMMA = Study of Muscle, Mobility and Aging; VO₂peak = peak oxygen consumption.

Performance deterioration was calculated as the inverse value of [(lap 2 − lap 9)/lap 2] × 100; higher values reflect greater slowing down (ie, greater performance fatigability).

^*Pearson’s (continuous) or Spearman (categorical) variables.

^†Normalized for intracranial volume.

^‡Normalized for total brain volume.

In models adjusted for age and sex, higher ¹¹[11C]-DTBZ binding in the sensorimotor striatum, but not in other striatal subregions, was significantly associated with lower performance deterioration (Table 3); this model explained 7% of the variance of the outcome. Results were similar in sensitivity analyses adjusted for overall speed during the fast 400 m walk. In logistic regression models with marked performance fatigability as the outcome, age and sex adjusted odds ratios were 0.30, 95% CI [0.09, 1.03], p = .056. Figure 1 illustrates the spatial distribution of the radioligand in 2 participants at the extremes of performance deterioration.

Table 3.

Associations of [¹¹C]-DTBZ Binding Potential in Striatal Subregions With Performance Deterioration as the Outcome

Independent Variables*	Outcome: Performance Deterioration
	1 Standard Deviation	Standardized β	95% Confidence Intervals
Posterior putamen	0.43	−1.08	−1.96, −0.20
Dorsolateral striatum	0.34	−0.76	−1.64, 0.12
Anterior ventral striatum	0.29	−0.66	−1.58, 0.25

Notes: Performance deterioration was calculated as the inverse value of [(lap 2 − lap 9)/lap 2] × 100; higher values reflect greater slowing down (ie, greater performance fatigability).

^*The [¹¹C]-DTBZ binding potential of each striatal subregion entered a separate model, adjusted for age and sex.

Figure 1.

Distribution of [¹¹C]-DTBZ binding potential in striatal subregions in 2 participants with higher and lower performance deterioration. Corresponding axial slices of 2 participants with higher (left) and lower (right) performance deterioration. Left: 80 years old woman with a 9.6% slower gait speed from lap 2 to lap 9. Right: 76 years old woman with a 2.5% slower gait speed from lap 2 to lap 9. The posterior putamen region is shown in the green outline. The hybrid images show a fusion of 3T MPRAGE structural MR image and a parametric image of [¹¹C]DTBZ binding potential (BP_ND) on a common scale ranging from 0 (lowest) to 4 (highest) in a single axial plane showing the corresponding volume of interest for posterior putamen. MR = magnetic resonance.

Adding maxOXPHOS and MoCA to the model did not substantially modify the association between ¹¹[11C]-DTBZ binding in the sensorimotor striatum and performance deterioration, and this model explained an additional 4% of the variance, for a total of 11% (Table 4). Further adjustment for other covariates did not modify the association of ¹¹[11C]-DTBZ binding in the sensorimotor striatum and performance deterioration, and explained an additional 10% of the variance for a total of 21%.

Table 4.

Association of [¹¹C]-DTBZ Binding Potential in the Posterior Putamen With Performance Deterioration

		Outcome: Performance Deterioration
		Standardized β	95% Confidence Interval	R 2 of the Model
Main Independent variable [11C]-DTBZ, Posterior Putamen (1 SD = 0.43)	Model 1	−1.02	−1.96, −0.09	0.11
	Model 2	−1.03	−1.93, −0.14	0.11
	Model 3	−1.13	−2.09, −0.18	0.21

Notes: All models are adjusted for age and gender.

Performance deterioration was calculated as the inverse value of [(lap 2 − lap 9)/lap 2] × 100; higher values reflect greater slowing down (ie, greater performance fatigability); SD = standard deviation. Model 1: further adjusted for Maximal Complex I&II Supported OXPHOS (pmol/(s*mg) + technician. Model 2: adjusted for Montreal Cognitive Assessment. Model 3: adjusted for Maximal Complex I&II Supported OXPHOS (pmol/(s*mg) + technician, Montreal Cognitive Assessment and VO2peak, mL/kg/min, body mass index, joint pain, medications.

Discussion

In this study of community-dwelling older adults, we found a distinct spatial distribution of higher ¹¹[11C]-DTBZ binding in the sensorimotor striatum in relation with lower performance deterioration, but not in other striatal subregions regulating reward or executive control function. Albeit statistically significant, ¹¹[11C]-DTBZ binding in the sensorimotor striatum explained a small portion of the variance of performance deterioration, and this proportion only marginally increased after accounting for other relevant predictors, such as muscle energetics or cognition. Our findings underscore the complexity of performance fatigability in older adults, and the need for a multisystem approach to better understand this potentially disabling condition. Intervention studies to ameliorate fatigability in older adults should consider the role of dopaminergic integrity in addition to muscle and cardiopulmonary systems.

The nigrostriatal dopaminergic system as a contributor to physical fatigability is rarely studied in older people without neurological diseases; a multisystem approach to fatigability in older age that includes central nervous system, skeletal muscle, and cardiopulmonary function is needed. While traditional models of fatigability have focused on muscle and cardiopulmonary factors, modern psychobiological models of exercise postulate that physical performance depends on factors within the central nervous system that balance perceived effort during exercise with the physiological ability to recruit motor neurons (31,33–36). In healthy young adults, increasing dopaminergic signaling reduces performance fatigability, dampening the perception of physiological markers of fatigability (rising heart and respiratory rate, muscle cramps), while increasing motivation and attention (6,7,37–39). Pharmacological approaches (such as levodopa, but also other agents) have been used in small samples of older adults, with improvements in higher-order cognitive function and depression (40–42), but its role in performance fatigability is unexplored. Future studies should assess the safety of pharmacological agents to restore dopaminergic neurotransmission in older adults without neurological diseases and the effects on fatigability.

Our results are consistent with prior work showing the relevance of dopamine for motor control in older adults without Parkinson’s disease or other neurologic disorders known to affect the dopaminergic system. We have shown that lower dopaminergic signaling (via PET (43–48), genetic polymorphism (49–51)) predicts worse physical function, independent of age and peripheral locomotor risk factors. In these studies, we found that higher dopaminergic signaling was associated with better mobility even among those with lower muscle strength, suggesting dopaminergic signaling may promote mobility, at least in part, via non-muscle pathways. Of note, muscle measures in this analysis did not modify the main association.

Contrary to our expectations, the association with dopamine signaling in other striatal subregions was not significant. We had hypothesized that dopamine in limbic and associative networks would influence performance deterioration, due to their role in regulating motivation and attention, important for motor control. Experiments increasing motivation and attention during high-intensity exercise show reduced perceived fatigability and improved performance (36,52). Age-related dopamine loss is associated with reduced reward-oriented behavior and working memory (53). It is possible that this task was not difficult enough to require motivation or executive control beyond the contribution of sensorimotor networks. A more challenging test requiring motivation and executive control, might have been more affected by dopamine binding in limbic and associative networks.

Measuring performance fatigability in real world setting in older adults is very challenging. A variety of tests have been developed, ranging from highly controlled laboratory settings to self-report perceptions. Although no gold standard currently exists for older adults, the measure of performance fatigability used in this study has been well represented in the literature, with good validity (15) and in settings resembling daily living conditions (54,55). However, a caveat of this metric is that it assumes a linear and consistent speed change from the 2nd to the 9th lap (56). Older adults may slow down in the middle of the walking task and speed up at the end, potentially because they may be motivated to finish. It is also possible that some participants walk slower than their capacity throughout the walk to avoid fatigue, or pain, or for other reasons; this self-pacing could be the result of an explicit decision-making process or an unconscious response to challenges. Our measure of fatigability may have missed these instances and would have classified these participants as having lower fatigability than their actual level, thus leading to an underestimation of their performance fatigability; these instances could explain the relatively small size of our estimates. Of note, performance fatigability did not appear to be influenced by motivation or decision-making processes in this study; the associations with dopaminergic integrity in reward and associative networks, perceived fatigability (a task requiring self-reflection of one’s capacities), depressive symptoms, or apathy were all not significant, thus it seems unlikely lower performance deterioration would have been largely due to self-pacing. Additionally, adjusting for overall speed during the walk did not modify the associations. Glynn and colleagues developed a more sensitive measure of performance deterioration using accelerometry-derived cadence over the entire walking task (57). Future studies should assess whether the role of dopamine to fatigability varies when using other metrics of performance fatigability.

The neuroimaging modalities used here are robust measures of dopamine integrity in the striatum. Neuroimaging values in this study are within the range of values previously published for adults of similar age indicating that our measures are consistent, and are thus comparable, across studies. [¹¹C]DTBZ is a widely used and well-characterized radiotracer for VMAT2; within-subject test-retest reliability is excellent at 3.6%–4.7% for caudate nucleus and putamen binding estimates. Participants enrolled in the present study were older adults with no history of any neurodegenerative disease and no documented use of a dopamine-enhancing drug such as those commonly prescribed in Parkinson disease patients. Therefore, it is highly unlikely that [¹¹C]DTBZ binding signal would be due to over- or downregulation of VMAT2.

We took great care to account for relevant contributors to fatigability. We have examined skeletal muscle mitochondrial capacity because it influences performance fatigability (5). MaxOXPHOS reflects maximal oxidative capacity of the mitochondria. The positive correlation between maxOXPHOS and performance deterioration was unexpected, and it may relate to the nature of the test, where some participants walked faster from lap 2 to lap 9, reflecting a pacing strategy that may not depend on muscle oxidative capacity. The influence of nonmeasured variables (fiber type proportion for example), or the walking efficiency of the participants may have also played a role. Another unexpected result was that cardiorespiratory fitness was not related with performance deterioration; we previously reported a significant association of lower VO₂peak with greater performance fatigability using a cadence-based measures in the parent cohort (4). Perhaps this selected subgroup of participants was too small to detect significant associations.

Limitations of this study should be noted. First, dopaminergic integrity was not measured during fatiguing tasks. However, our protocol provides valid and reliable measures of dopaminergic integrity, with demonstrated associations with physical and cognitive measures outside the scanner. Since this is a cross-sectional analysis, descriptive associations cannot capture mechanisms underlying fatigability. This cohort displayed a relatively small performance deterioration on average, indicating adults participating in this study are on the high-functioning end of the spectrum and our results may not extend to the general population of older adults. Of note, the prevalence of marked performance fatigability, defined as performance deterioration of ≥6.5%, was similar to that observed in other cohort of similarly aged older adults living in the community (14,15). Given the relatively small performance deterioration observed in this cohort, it is remarkable that there was a statistically significant association between dopamine integrity and performance deterioration. Strengths of this study include the comprehensive characterization of multiple systems potentially influencing performance fatigability, and also the focus on older age. Previous studies of neurobiological aspects of fatigability examined much younger individuals, and often used less sensitive measurements of fatigability.

In sum, our results indicate potentially modifiable neural markers of performance fatigability, a key first step toward a mechanistic understanding of neural processes contributing to optimal performance, and toward novel multisystem interventions to reduce age-related fatigability and its disastrous sequelae.

Funding

The Study of Muscle, Mobility and Aging is supported by funding from the National Institute on Aging (R01 AG 059416). Study infrastructure support was funded in part by NIA Claude D. Pepper Older American Independence Centers, at University of Pittsburgh (P30 AG024827) and Wake Forest University (P30 AG021332) and the Clinical and Translational Science Institutes, funded by the National Center for Advancing Translational Science, at Wake Forest University (UL1 TR001420). This analysis was supported by NIA awards (R01AG075025 and U01AG061393).

Conflict of Interest

None.