A person’s mental state is an important intrinsic factor affecting the expression of a variety of neurological motor symptoms, from dystonia to hemiparesis after stroke [1]. Parkinson’s disease (PD) is a prototypical example highlighting this captivating interaction, where alterations in the mental state often lead to immediate observable changes in motor symptoms. While tremor is the prime example of a PD motor feature that typically worsens during stressful situations and with increased cognitive load [2,3,4,5,6], other motor symptoms, such as gait impairments, are also often aggravated by stress and anxiety [6,7,8]. Gait impairments in PD comprise both continuously present deficits (e.g., reduced step length and height, reduced gait speed, and increased gait variability) as well as episodic deficits (i.e., festination and freezing of gait) [9]. The detrimental effect of stressful, anxiety-inducing situations on the occurrence and severity of freezing of gait episodes has been particularly well-established [10,11,12,13,14,15,16]. Unsurprisingly, many people with PD and gait impairment use strategies to counteract stress or anxiety, in an effort to improve their walking ability. Of these approaches, mindfulness-based interventions have received the most attention in recent years [17,18,19,20].

Paradoxically, in contrast to this negative impact on gait, some PD patients seem to actually benefit from experiencing—or even purposely inducing—stressful situations. The most extreme form is the well-known phenomenon of ‘kinesia paradoxa’: a sudden, transient ability to perform a task that a person was previously unable to complete, often in the context of grave, life-threatening situations [21,22]. A classic case example was offered by a group of 14 institutionalized Italian patients with advanced PD, who demonstrated an extraordinary motor improvement during the L’Aquila earthquake of 2009 [23]. All of them were able to escape unaided from the collapsing nursing home, despite usually requiring assistance during daily activities because of severe gait difficulties and postural instability. While kinesia paradoxa may well be an entity on its own, it has been hypothesized that gait improvement under more mundane circumstances that increase motivation may share a similar underlying mechanism, namely a shift to optimally heighten arousal [24]. These common clinical observations raise an interesting question—how is it that heightened arousal can augment symptoms in some individuals, while alleviating their severity in others?

Here, we suggest that this paradox is related to inherent features of the ascending arousal system. Over a century ago, psychologists Yerkes and Dodson demonstrated that task performance is related to arousal (Fig. 1) [25]. The inverted u-shaped curve illustrates that task performance improves with increasing arousal, until an optimum is reached. When the optimal level of arousal is surpassed, performance declines again as arousal increases further. Depending on where you are situated on this curve, a change in arousal could therefore be either beneficial or detrimental to performance of the task at hand. Translating this notion to gait impairments in PD, on the left-hand side of this curve (i.e., with suboptimal arousal), gait might be ameliorated by strategies that increase arousal levels. Conversely, on the right-hand side of the curve (i.e., with supraoptimal arousal) gait might benefit from strategies that decrease arousal.

Compensation strategies tapping into this mechanism of modulating arousal levels are typically attributed to a category entitled ‘Altering the Mental State’. This is one of seven major categories of compensation strategies for gait impairment, based on a comprehensive review of several hundred videos of persons with PD who spontaneously ‘invented’ strategies to improve their gait [24]. The ‘Altering the Mental State’ category has thus far received relatively little attention, certainly when compared to other categories such as external or internal cueing [26]. For obvious reasons, it is not recommended to implement life-threatening situations in daily life to improve gait impairment. However, reports on ‘everyday’ variants of less drastic arousal strategies are scarce. Indeed, PD healthcare professionals rarely apply strategies from the ‘Altering the Mental State’ category in their clinical practice [27], even though persons with PD find it a valuable way to cope with their gait impairments in daily life [28]. Furthermore, the mechanisms underlying this category of strategies remain poorly understood.

Here, we introduce a novel theoretical framework regarding the potential underlying mechanisms of modulating one’s arousal to optimize gait performance in PD. To this aim, we first present an array of ready-to-use ‘Altering the Mental State’ strategies for gait impairment, informed by data from a large-scale international survey study among over 4000 persons with PD and gait impairments [28], illustrating our hypothesis that the noradrenergic system may act as a double-edged sword in PD gait control.

Here, we present an overview of ‘Altering the Mental State’ strategies that persons with PD applied in an effort to overcome their gait impairments in daily life. Approximately one in three patients with gait impairments reported altering their mental state as a compensatory strategy. While most persons with PD reported to have used strategies that reduced arousal (e.g., by diminishing stress, or promoting general relaxation), a smaller group actually reported to have purposely increased their arousal to improve gait. We will next elaborate on our theoretical framework regarding the potential underlying mechanisms and address the clinical implications of self-modulating arousal to optimize gait performance in PD.

From a neurobiological perspective, the concept of fine-tuning one’s arousal level to optimizing one’s ability to compensate for gait impairments in PD is tightly connected to the functions of the locus coeruleus. The locus coeruleus is a small nucleus located in the posterior area of the rostral pons, and represents the primary source of noradrenaline for the central nervous system [31]. Its widespread noradrenergic projections modulate cortical, subcortical, cerebellar, brainstem and spinal cord circuits, which makes it well suited to rapidly and globally modulate brain function in response to changes in the environment (e.g., stressful stimuli) [32]. Moreover, locus coeruleus noradrenaline contributes to the reconfiguration of functional communication between distributed brain regions [33]. It is part of the ascending arousal pathway, and plays a major role in attentional and arousal response to threat [34]. Indeed, locus coeruleus activity displays an inverted u-shaped relationship with task performance, in accordance with Yerkes-Dodson’s law [35,36]. Therefore, changes in the firing rate of the locus coeruleus may facilitate a shift in arousal that promotes optimal performance on the task at hand (Fig. 1). The neuromodulatory impact of the locus coeruleus on the central nervous system is classically compared to tuning the volume of a radio, even though its precise adaptive role is more complex and dynamic, and more appropriately analogous to the bowing of a violin [37]. Much like the effect that a bow has on the strings of a violin, the locus coeruleus changes the way the ‘notes’ are expressed (e.g., their volume or tone quality) without affecting the specific string of notes in itself, in order to shape complex neuronal melodies [37]. Increased activity of the locus coeruleus (i.e., turning up the volume) increases the strength of functional interactions between brain regions that are otherwise largely segregated [38,39], as noradrenaline elicits changes in the internal milieu of target cells to alter their ‘neural gain’ (i.e., their excitability and receptivity to incoming signals) [40]. By mediating this increased network-level integration across the brain, the locus coeruleus can facilitate dynamic ‘cross-talk’ between different regions across the cortex and subcortex that are critical for higher-order functions [41,42]. This functional integration is of particular importance in gait control in PD.

The pathophysiology underlying gait impairments in PD is complex and presumably involves dysfunction of multiple cortical and subcortical components. Gait partly depends on a basic ‘locomotor network’, involving spinal central pattern generators, brainstem mesencephalic and cerebellar locomotor regions, along with the corticostriatal input projecting to the primary motor cortex [43,44]. In addition, distributed cortical areas, particularly the frontoparietal and supplementary motor areas, are normally involved in the adjustment and adaptation of walking [45]. During walking in an automated manner (i.e., without consciously paying attention to it), persons with PD typically have difficulties recruiting these cortical motor areas [46]. Adequate gait control therefore not only relies on the integrity and function of corticostriatal motor loops, but on compensatory input from cognitive, sensory and limbic systems as well (Fig. 2) [47,48,49,50]. Reinforcing the integration of these different neural networks may consequently facilitate optimal compensation for the PD-related loss of function in the motor circuitries. When locus coeruleus activity is too low (i.e., suboptimal arousal), the different compensatory networks may be too segregated to allow for adequate compensation of motor impairments. Overactivation of the locus coeruleus (i.e., supraoptimal arousal), on the other hand, may lead to a situation that allows for an element of increased, dysfunctional ‘cross-talk’ between competing inputs of these complementary networks [51]. In PD specifically, this has been associated with detrimental effects on motor function, particularly the occurrence of anxiety-induced freezing of gait [52,53,54]. Unfortunately, the ability to adaptively employ the locus coeruleus to optimally modulate the interaction of compensatory networks may be affected by the profound degeneration of the nucleus in PD, and this is compounded by dysfunction of cortical regions in charge of regulating arousal. Whilst well-established, this loss of noradrenergic neurons (ranging 20%-90% in PD patients [55,56]) has been relatively neglected [57], even though it both precedes the onset and exceeds the extent of the hallmark loss of dopaminergic neurons in the substantia nigra pars compacta [58,59,60]. The application of ‘Altering the Mental State’ strategies to modulate arousal may therefore be necessary to facilitate shifts from sub- or supraoptimal states of arousal to optimize gait performance in PD.

Importantly, the underpinnings of these strategies presumably do not only involve noradrenaline, but other neurotransmitters such as dopamine and serotonin as well [61]. The classic parallel model suggesting a direct correlation between changes in a single neurochemical system and a distinctive deficit is likely to be too simplistic. Rather, a convergent biochemical model in which the complex interactions between the different monoaminergic systems are taken into account would be more appropriate [62]. For example, the concept of reaching a “flow state”, which respondents reported to be helpful in ameliorating gait, illustrates that striking the right balance between motivation (dopamine) and arousal (noradrenaline) can lead to optimized behavior [63,64]. Arousal levels are also influenced by a multitude of both internal and external factors (e.g., one’s present emotions or physical environment). Even the presence of gait impairments in itself is likely to influence arousal levels in people with PD: it introduces a vicious cycle of supraoptimal arousal due to anxiety related to falling, in turn leading to aggravated gait impairment, which further increases the fear of falling. However, while a person’s ‘baseline’ position on the curve is likely to be partly trait-dependent (e.g., a chronic state of supraoptimal arousal in those with generalized anxiety or a debilitating fear of falling), it would presumably be predominantly dependent on the task, environment and context at hand (e.g., a sudden increase in arousal when a doorbell rings). In light of this, it is unsurprising that a proportion of respondents reported to have used both types of ‘Altering the Mental State’ strategies in daily life. One can imagine that depending on the situation at hand, the state of arousal—and with that the appropriate type of strategy—may vary. Furthermore, depending on the complexity of the (gait) task at hand, the optimal level of arousal (i.e., the shape of the curve) could vary significantly across tasks [25,65,66]. These factors combined may make it particularly difficult to employ pharmacological interventions targeting the noradrenergic system as an add-on to dopaminergic medication to improve gait performance in persons with PD [67,68,69,70]. Any medication targeting the noradrenergic system would potentially impact the trait-dependent ‘baseline’ arousal level, rather than enabling the necessary dynamic adaptation in arousal that might be required over the course of the day. An advantage of the application of non-pharmacological compensation strategies is that it allows for such dynamic approaches that are tailored to the specific personal needs of individual patients under everchanging circumstances. Future studies could investigate whether a more personalized approach to the use of pharmacological agents targeting arousal could be beneficial in selected groups of patients (e.g., those with particularly high or low trait-dependent anxiety levels as measured by existing scales like the State-Trait Anxiety Inventory). For example, if one occupies a highly aroused state the majority of their day, could ‘resetting’ their baseline arousal level help shift them to occupy an optimal part of the curve for a greater proportion of their daily activities? Quantifying baseline levels of locus coeruleus activity using techniques such as neuromelanin-sensitive MRI [71,72] may hold promise to stratify patients into clinical trials according to their level of noradrenergic (dys)function [73].

The survey data that we used to support our hypothesis rely on the self-reported presence of a PD diagnosis, presence of gait impairment, and the efficacy of the applied mental strategies. The potential presence of selection bias, given the selected and rather proactive subpopulation of PD patients that is typically involved in online cohorts, must also be taken into account when interpreting the survey data. However, these data were merely used to probe the presence of examples of strategies at two ends of the arousal spectrum, rather than to provide an accurate percentage of the prevalence of specific strategies. Nevertheless, future clinical trials are necessary to objectively quantify the efficacy of ‘Altering the Mental State’ strategies in improving gait performance in PD. In conjunction, measuring the effect of applying these strategies on physiological markers of arousal (e.g., measured by skin conductance, heart rate, pupil diameter) would help establish the plausibility of our proposed neurobiological framework. Lastly, looking into specific patient characteristics that may be associated with the efficacy of either arousal-reducing (e.g., high levels of trait anxiety) or arousal-increasing strategies (e.g., high levels of apathy or anhedonia) will be essential to eventually work towards a more personalized approach to the use of these non-pharmacological strategies in clinical practice [74]. Indeed, a subgroup characterization of the present cohort revealed that women less frequently reported to have used strategies to increase arousal, which may be a reflection of the higher prevalence of anxiety among women with PD [75].

Our theoretical framework proposing a central role for the locus coeruleus in facilitating optimal compensation to address gait impairments provides future opportunities to investigate the control of gait, as well as guiding both targeted pharmacological (e.g., tailored use of noradrenergic agents as an add-on to dopaminergic medication in selected PD patients) and non-pharmacological therapies (e.g., a training program on ‘Altering the Mental State’ strategies for persons with PD). It is also possible that the impaired modulation of arousal arising from the dysfunction of the locus coeruleus may have consequences that extend beyond gait and could play a critical role in PD tremor, ‘wearing off’ periods, or other key symptoms that have been demonstrated to be significantly affected by one’s level of arousal (particularly in the context of stress or anxiety) [6]. Future work is necessary to validate this conceptual framework, quantify the efficacy of ‘Altering the Mental State’ strategies, and further crystallize the potential involvement of the noradrenergic system in optimizing motor performance in PD.

PD:Parkinson’s disease

The Fox Insight Study is funded by The Michael J. Fox Foundation for Parkinson’s Research. We would like to thank the Parkinson’s community for participating in this study to make this research possible. The authors would also like to thank ParkinsonNEXT for facilitating the distribution of the survey in the Netherlands. This study was supported by a ZonMW Veni grant to Jorik Nonnekes (16.196.022), and an Academy Van Leersum grant of the Academy Medical Sciences Fund, Royal Netherlands Academy of Arts & Sciences to Anouk Tosserams. The Center of Expertise for Parkinson & Movement Disorders was supported by a center of excellence grant awarded by the Parkinson Foundation.

AT: concept and design of the study, data collection, interpretation of data, drafting the manuscript. BRB: concept and design of the study, interpretation of data, critical revision of the manuscript. KAEM: interpretation of data, critical revision of the manuscript. RCH: interpretation of data, critical revision of the manuscript. RPCK: interpretation of data, critical revision of the manuscript. JMS: interpretation of data, critical revision of the manuscript. NLT: interpretation of data, critical revision of the manuscript. GW: interpretation of data, critical revision of the manuscript. SJGL: interpretation of data, critical revision of the manuscript. JN: concept and design of the study, data collection, interpretation of data, critical revision of the manuscript. All authors read and approved the final manuscript.

AT was supported by an Academy Van Leersum grant of the Academy Medical Sciences Fund, Royal Netherlands Academy of Arts & Sciences. BRB currently serves as Editor-in-Chief for the Journal of Parkinson’s disease, serves on the editorial board of Practical Neurology and Digital Biomarkers, has received honoraria from serving on the scientific advisory board for the Critical Path Institute, Kyowa Kirin, UCB and Zambon, has received fees for speaking at conferences from AbbVie, Bial, GE Healthcare, Oruen, Roche and Zambon, and has received research support from the Netherlands Organization for Scientific Research, the Michael J Fox Foundation, Nothing Impossible, Parkinson Vereniging, the Parkinson’s Foundation, Hersenstichting Nederland, Verily Life Sciences, Horizon 2020, the Topsector Life Sciences and Health, Abbvie, UCB, Yesse Technologies and Zambon, all outside the submitted work. KAEM was supported by a Parkinson Canada New Investigator grant, as well as a Natural Sciences and Engineering Research Council of Canada Discovery grant. RCH was supported by a research grant (VIDI; #09150172010044) from the Netherlands Organization for Scientific Research. He has acted as consultant for UCB (unrelated to this work). RPCK was partially supported by a grant from the European Regional Development Fund (ERDF/EFRO, grant number PROJ-00928) outside the submitted work. He is associate editor for Neuropsychology Review, member of the editorial board of the Journal of the International Neuropsychological Society, member of the scientific advisor board of Alzheimer Nederland, chair of the scientific advisory board of the Korsakoff Knowledge Center Netherlands (Korsakov Kenniscentrum), and scientific consultant for the University of Michigan. JMS was supported by the National Health and Medical Research Council (1193857). NLT was supported by the Australian Government Research Training Program (RTP) Scholarship. GW reports no disclosures. SJGL was supported by a National Health and Medical Research Council Leadership Fellowship (1195830). JN was supported by a ZonMW Veni grant (16.196.022). He reports a grant from ZonMW (OffRoad grant), Michael J. Fox Foundation, Ipsen Pharmaceuticals and Gossweiler Foundation outside the submitted work.

All individual data collected through the Fox Insight online clinical study are available on the Fox Insight database (https://foxinsight-info.michaeljfox.org/insight/explore/insight.jsp). Qualified researchers, interested in PD or related research, can access Fox Insight data upon account registration, completion of a Data Use Agreement (DUA), acknowledgement of the study publication policy, and review by the Data and Publications Committee. Other data related to the present work are available upon request to the corresponding author.