Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Oct 3.
Published in final edited form as: Exp Brain Res. 2024 Oct 3;242(12):2853–2863. doi: 10.1007/s00221-024-06933-5

Blending motor learning approaches for short-term adjustments to gait in people with Parkinson disease

Chelsea Parker Duppen a, Nikhil Sachdeva b,c, Hailey Wrona d, Eran Dayan e, Nina Browner f, Michael D Lewek a,g
PMCID: PMC12372860  NIHMSID: NIHMS2094727  PMID: 39361030

Abstract

Rhythmic auditory cueing (RAC) using an isochronous metronome is an effective approach to immediately enhance spatiotemporal aspects of gait for people with Parkinson disease (PwPD). Whereas entraining to RAC typically occurs subconsciously via cerebellar pathways, the use of metronome frequencies that deviate from one’s typical cadence, such as those used in rehabilitation, may require conscious awareness. This heightened awareness may increase cognitive load and limit the persistence of gait training gains. Here, we explore the immediate effects of incorporating an implicit motor learning approach (i.e., error-based recalibration) to gait training with RAC. Twenty older adults (10 with PD and 10 controls) were asked to match their footfalls to both isochronous and subtly varying metronomes while walking on a treadmill and overground. Our findings revealed intriguing differences between treadmill and overground walking. During treadmill walking to a slower metronome frequency, both groups reduced their cadence and increased step lengths, but did not make the necessary adjustments to match the subtly varying metronome. During overground walking, both groups modified their cadence in response to a 3–4% change in metronome frequency (p < 0.05). Both metronomes yielded evidence of implicit and explicit retention during overground and treadmill walking. Furthermore, during overground walking the PD group showed greater implicit retention of cadence changes following the varying metronome, compared to the isochronous metronome. Our results suggest that incorporating implicit motor learning approaches to gait training during a single session of overground walking may enhance short term implicit retention of gait behaviors for PwPD.

Keywords: Motor Learning, Gait, Parkinson Disease, Rehabilitation, Neurorehabilitation

Introduction

People with Parkinson disease (PwPD) typically walk with reduced step lengths, and slower gait speed than their similarly-aged, unimpaired peers (Morris et al. 1994; Azulay et al. 1999; Galna et al. 2015). Rhythmic auditory cueing (RAC) is widely recognized as an effective strategy to immediately improve these spatiotemporal aspects of gait in PwPD (Willems et al. 2006; Nieuwboer et al. 2007; Ashoori et al. 2015; Ghai et al. 2018; Thaut et al. 2019; Hoppe et al. 2020; Sherron et al. 2020). During RAC, participants entrain to external rhythmic cues, such as music or a metronome, which helps overcome impaired internal rhythm generation due to dopaminergic loss (Puyjarinet et al. 2019). Although entrainment to RAC is believed to occur subconsciously via cerebellar pathways (Del Olmo et al. 2007; Fujioka et al. 2012), researchers and clinicians may inadvertently induce conscious awareness when employing metronome cue frequencies that are dramatically different than typical cadence, such as those used to increase gait speed and/or step length during training (Willems et al. 2006; Hoppe et al. 2020; Sherron et al. 2020). The potential for conscious awareness during training is suggestive of a strategic learning approach, which is an explicit form of motor learning and may be cognitively taxing (French et al. 2021). Given that PwPD demonstrate difficulty with cognitive-motor dual tasks (O’Shea et al. 2002; Raffegeau et al. 2019), this conscious involvement may contribute to the short-lasting benefits of RAC, with declines in performance observed as little as two months after training (Nieuwboer et al. 2007; Thaut et al. 2019).

The inclusion of implicit motor learning mechanisms, such as error-based recalibration (Mazzoni and Krakauer 2006; Taylor et al. 2014; Long et al. 2016; Leow et al. 2018; Duppen et al. 2023), are associated with improved retention of learned motor behaviors compared to explicit-mediated approaches (Lee and Vakoch 1996; Kim et al. 2015). Error-based recalibration involves adjusting movement patterns based on the difference between anticipated and actual performance of a task. By detecting and correcting these discrepancies through practice, individuals refine their motor control. This adaptation can be triggered by subtly changing the context in which a task is being completed, prompting people to adapt their movements accordingly. The ability of individuals to adapt movement rhythm implicitly when rhythmic cues are subtly varied is well-supported in the literature during seated lower extremity tasks (Chen et al. 2006), standing lower extremity tasks (Chen et al. 2006), and finger tapping (Repp 2000). Moreover, successful adaptation of gait, a complex motor pattern, has been observed in young, unimpaired adults who were asked to match their footfalls to a subtly varying metronome (Forner-Cordero et al. 2019; Duppen et al. 2023).

The use of error-based recalibration, in contrast to other implicit motor learning mechanisms, holds particular promise for PwPD. Error-based recalibration via a subtly varying metronome necessitates continuous rapid encoding, which engages the hippocampus – a neural structure largely preserved in PwPD (Henke 2010; Jablonowski et al. 2018). In contrast, other implicit motor learning mechanisms, such as use-dependent learning, rely on proceduralization, which utilizes striatal pathways impacted by dopaminergic loss in PD (Henke 2010; Jablonowski et al. 2018; Owens-Walton et al. 2018). It remains unclear, however, whether individuals with altered neural control, such as those with PD, can effectively adjust spatiotemporal aspects of their gait in response to a subtly varying metronome.

The purpose of this study is to build on prior work involving the use of subtly varying metronomes to incorporate implicit motor learning approaches into rehabilitation interventions. Specifically, we sought to investigate the ability of naïve older adults with and without PD to adjust their cadence and step lengths in response to a subtly varying metronome while walking on a treadmill and overground. Furthermore, we aimed to investigate the contributions of implicit and explicit motor learning when walking to a constant (i.e., isochronous) and subtly varying metronome, focusing on cadence, step length, and gait speed. Building off prior work in young, unimpaired populations (Forner-Cordero et al. 2019; Duppen et al. 2023), we hypothesized that subtle variations in metronome cue frequency would elicit adjustments to stepping cadence during both treadmill and overground walking for older adults and PwPD. Additionally, we hypothesized that a subtly varying metronome would lead to greater implicit retention of practiced gait patterns than the isochronous metronome for both groups, due to the addition of error-based recalibration.

Materials and Methods

Participants

We recruited people with idiopathic PD and unimpaired similarly aged adults (Controls) to complete treadmill and overground testing. Participants were included if they were between the ages of 50 and 99 years old, able to walk overground without physical assistance or an assistive device for ≥ five minutes, and were able to walk on the treadmill at ≥ 80% of their comfortable overground gait speed. Participants with PD were included if they had a diagnosis of idiopathic PD from a neurologist and fell within stages I-III on the Hoehn and Yahr scale. Participants were excluded if they had a deep brain stimulator, Parkinson-related or unrelated dementia (Montreal Cognitive Assessment [MOCA] score < 23) (Hsu et al. 2015), other neurologic or orthopedic disorders that may affect gait, or severe communication impairments. All participants signed an informed consent form approved by the Institutional Review Board of the University of North Carolina at Chapel Hill (IRB# 20–3099).

Overall Design

Prior to testing, all participants completed the MOCA (Hsu et al. 2015) and Mini-BESTest (King et al. 2012). Next, we employed a randomized crossover design in which participants completed two sessions (isochronous vs varying frequency metronome) separated by at least seven days. We have described these methods previously (Duppen et al. 2023). Briefly, however, participants were randomly assigned to receive either the isochronous or subtly varying metronome during the initial session with the other metronome used during the second session. Importantly, participants were not told that the metronome frequency would change during one of the two sessions.

Initially, we conducted baseline testing to determine participants’ comfortable overground gait speed and cadence using two passes across a 4.27m (14 ft) GAITRite mat (CIR systems, Franklin, NJ). We used the cadence values from baseline testing to set the metronome frequency, and the comfortable gait speed to set treadmill (Bertec, Worthington, OH) speed. Following baseline overground testing, participants walked on a treadmill that was set to their comfortable overground gait speed for one minute without any cueing. If participants were unable to walk on the treadmill at their comfortable overground gait speed, they were asked to walk at their fastest attainable speed, so long as it was > 80% of their comfortable overground speed. Participants were then asked to walk on the treadmill while matching their footfalls with a metronome for 9 minutes (see details below regarding the isochronous and varying metronomes). After the 9-minute trial, we turned off the metronome and assessed for implicit and explicit retention effects. Participants were asked to first “walk normally” (implicit retention) and then “walk how you feel the metronome was having you walk” (explicit retention) for 30 seconds each. This pattern of testing was then repeated for overground walking.

Isochronous metronome session:

For treadmill testing, the metronome was set to 85% of each participant’s comfortable overground cadence to elicit increased step lengths (Hoppe et al. 2020). During overground testing, the metronome was set to 115% of each participant’s comfortable overground cadence to elicit increased gait speed (Hoppe et al. 2020; Sherron et al. 2020). We used targeted rhythmic auditory cueing (TRAC) as opposed to a more conventional RAC (i.e., cues that are 100% of typical cadence) based on its ability to impact spatial and temporal aspects of gait for people with PD (Hoppe et al. 2020; Sherron et al. 2020).

Varying metronome session:

During treadmill testing the metronome began at 85% of each participant’s comfortable overground cadence for one minute. The metronome then subtly increased by 1 percent of comfortable cadence every 20 seconds until it reached 91%. Following the 91% cue, the metronome then decreased by 1% every 20 seconds back to 85% of typical overground cadence. This pattern was repeated for a total of 9 minutes to allow the cycle to run twice. For overground testing, the metronome varied from 115-109-115% of baseline overground cadence, again changing by only one percent every 20 seconds (Figure 1). To minimize the potential benefits of the varying metronome beyond the need for constant recalibration of movements, we intentionally set the extremes of each metronome to be identical (i.e., the lowest frequency of the varying metronome on the treadmill was 85%, and the highest frequency of the varying metronome overground was 115%). Thus, the average frequency of the varying metronome was slightly higher on the treadmill and slightly lower overground compared to the isochronous metronome, ensuring that the varying condition did not receive more extreme ranges.

Fig. 1.

Fig. 1

Open in a new tab

Schematic of study design demonstrating target cadence for each metronome type for both treadmill and overground walking. Adapted from Duppen et al (Duppen et al. 2023)

After the second testing session was completed, we asked participants “Did you notice a difference between the two metronomes?” If a participant responded ‘yes’ and correctly identified that one of the metronomes was changing frequency, their response was recorded as a ‘yes’. Conversely, if a participant either responded ‘no’ or responded, ‘yes’ but provided an incorrect explanation (e.g., one of the metronomes had a different pitch, etc.), their response was recorded as a ‘no’. This helped determine if the subtle variations in metronome frequency remained below the threshold of conscious awareness during the varying condition.

Data Collection

During treadmill testing, participants walked on a dual-belt instrumented treadmill (Bertec, Worthington, OH). Ground reaction forces were recorded at 960Hz from the treadmill, and heel position data were obtained at 120Hz via 14-mm retroreflective markers attached to participants’ heels using an 8-camera motion capture system (Vicon, Los Angeles, CA). For safety, participants wore an overhead harness that did not provide bodyweight assistance or restrict movement. Although we encouraged participants to avoid handrails, if participants required their use, handrails were used only for balance support and maintained consistent across both sessions.

Overground testing was completed in a 100’ hallway. As participants walked during the 9-minute metronome conditions, we recorded video of participants from the waist down to allow for cadence calculations. During the overground baseline, implicit, and explicit retention testing participants performed four passes across the GAITRite mat for each condition.

Data Management and Analysis

Marker trajectories and ground reaction forces during treadmill testing were filtered with a fourth-order low-pass 6 and 25 Hz Butterworth filter, respectively. Heel markers were used to calculate step lengths as the distance between feet in the anterior direction at each heel strike. We used ground reaction forces to determine heel strikes and calculated cadence as the inverse of step times. For the overground baseline, implicit, and explicit conditions, we used PKMAS software (Protokinetics, Havertown, PA) to calculate cadence, gait speed, and step length. During the 9 minutes of overground walking to a metronome (isochronous and varying), we determined cadence by counting the number of steps taken within each 20-second time period based on video data, corresponding to the changes in metronome frequency. All outcome measures (i.e., step length, cadence, and gait speed) were normalized to each participant’s original overground baseline trial.

Data were analyzed using SPSS (v28; IBM Corp, Armonk, NY). We used a mixed-design analysis of variance to examine the main and interaction effects of group (i.e., PwPD or controls), session (i.e., isochronous or varying), and time period on cadence and step length (treadmill only). This approach was chosen to allow for the examination of both within-subjects factors (session and time period) and between-subjects factors (group). Time periods were combined across similar metronome cues and the corresponding time during the isochronous cue (termed ‘epoch’ here). When significant main and interaction effects were observed, we carried out planned one and two-way ANOVAs and paired sample t-tests for post-hoc comparisons.

To investigate implicit and explicit retention effects, we completed a mixed-design repeated measures ANOVA to compare baseline, implicit, and explicit performance during the constant and varying conditions between groups (i.e., PwPD or Controls). When significant main and interaction effects were observed, we again carried out planned one and two-way ANOVAs and paired sample t-tests for post-hoc testing.

We used an alpha = 0.05 to determine levels of significance. Effect sizes are reported as partial eta squares and interpreted using standard benchmarks to indicate small, medium, or large effects (Richardson 2011). To control for the increased risk for Type I error due to multiple comparisons, we applied Bonferroni corrections during post-hoc testing.

Results

Ten unimpaired older adults (68 ± 8 years of age; 5 females, 5 males) and twelve PwPD (66 ± 11 years of age; 5 females, 7 males) participated. Two PwPD were not included in the analysis due to an inability to complete both sessions for reasons not related to this study. Of the participants who completed both testing sessions, there was no difference in age or sex between the older adult controls and individuals with PD (p = 0.663 and p = 0.991, respectively). There was also no difference in comfortable overground gait speed, cadence, or step length (all p > 0.1) (Table 1). For PwPD, four were Hoehn & Yahr Stage 3, two were classified as Stage 2, and four were Stage 1. Nine out of ten participants in each group were unaware of the difference between metronome conditions (isochronous vs varying). The participant with PD able to detect the change in metronome is denoted by an open circle on Figures 2 and 4.

Table 1:

Participant Characteristics (shown as mean ± standard deviation)

PwPD Group Controls p-value
Age (years) 66 ± 11 68 ± 7 0.663
Sex 4F, 6M 5F, 5M 0.991
Mini-BESTest Score 26 ± 2 27 ±1 0.07
MOCA 28 ± 2 28 ± 2 0.480
UPDRS (total) 26 ± 12 N/A N/A
Gait Speed (m/s) 1.09 ± 0.18 1.22 ± 0.18 0.166
Step Length (m) 0.617 ± 0.080 0.655 ± 0.065 0.283
Cadence (steps/min) 105 ± 8 108 ± 9 0.476
Open in a new tab

Fig. 2.

Fig. 2

Open in a new tab

(A) Percent comfortable overground cadence and step length demonstrated by each group while walking to the subtly varying metronome on the treadmill. The average for PwPD is outlined by diamonds and a thick red line while the control group average is denoted by squares and a thinner black line with the standard deviation shaded in blue. Individual data points for each epoch are shown for the ten participants with Parkinson disease. (B) Percent comfortable overground cadence and step length demonstrated by each group during treadmill walking with the isochronous metronome. For both (A) and (B), the participant with PD who was able to state that one metronome was varying is denoted by an open pink circle

Fig. 4.

Fig. 4

Open in a new tab

Relative cadence for each group while walking overground to the (A) subtly varying and (B) isochronous metronomes. Cadence is represented by percent of comfortable overground cadence. The control group is represented by the thin black line and square shapes for average cadence with one standard deviation from the mean represented by light blue shading. Individual data points represent cadence of each participant with PD during the time period, with the average of PwPD noted by a thicker red line. The participant who was able to detect that one metronome was varying is denoted by an open pink circle on the graphs

Treadmill Testing

During treadmill testing, cadence was seen to have a significant main effect of group (p < 0.001, ηp2 = 0.027) and session (p < 0.001, ηp2 = 0.023), with no main effect of epoch (p = 0.879, ηp2 = 0.002). Additionally, there was an interaction effect of group × session (p = 0.002, ηp2 = 0.010; Figure 2a). In particular, we observed that participants with PD demonstrated a lower relative cadence during the isochronous condition as compared to the control group (p < 0.001), whereas there was no difference between the two groups during the varying condition (p = 0.110). Across both groups, cadence was lower in the isochronous session than the varying session (p = 0.003). For step length, we found a significant main effect of condition (p < 0.001, ηp2 = 0.034), such that step lengths were longer during treadmill walking with the isochronous metronome. No main effects of group (p = 0.143, ηp2 = 0.002) or epoch (p = 0.971, ηp2 = 0.001) were found for step length during treadmill testing. However, there was an interaction effect of group × condition (p = 0.018, ηp2 = 0.006; Figure 2b). Specifically, we observed that participants with PD took shorter step lengths during the varying condition than older adult controls (p = 0.005), whereas there was no difference between the two groups during the isochronous condition (p = 0.425).

We assessed for implicit and explicit retention after each treadmill condition. In comparing the baseline, implicit, and explicit retention, we observed a significant main effect of cue only (i.e., baseline, “walk normally”, or “walk how you feel the metronome was having you walk”) for both cadence (p < 0.001, ηp2 = 0.761) and step length (p < 0.001, ηp2 = 0.791), such that the implicit and explicit retention cues resulted in decreased cadence and increased step length compared to the baseline, and that the explicit retention cue resulted in lower cadence and increased step length compared to the implicit retention testing (all p < 0.001) (Figure 3). Importantly, there were no main effects of group or condition, nor were there interaction effects of group × condition or cue × condition for cadence or step length (all p > 0.05).

Fig. 3.

Fig. 3

Open in a new tab

Implicit and explicit retention effects of (A) cadence and (B) step length after stepping to the isochronous (open shape) and subtly varying (closed shape) metronomes on the treadmill. Cadence and step length are represented by the percent of comfortable overground cadence and step length, respectively. Baseline represents how they walked on the treadmill prior to introduction of either metronome. Each participant’s average cadence and step length are represented by a data point on the figure

Overground Testing

During overground walking, we observed main effects of condition (p < 0.001, ηp2 = 0.013) and epoch (p < 0.001, ηp2, = 0.058) as well as interaction effects of group × condition (p = 0.001, ηp2 = 0.020) and condition × epoch (p < 0.001, ηp2 = 0.059) (Figure 4). Specifically, we found that a change of 3–4% of the metronome frequency was sufficient to observe a change in cadence during the varying condition for both groups (p < 0.05), whereas there was no change in cadence during the constant condition (p = 0.961). Furthermore, we observed that PwPD did not demonstrate a difference in cadence between the two conditions (p = 0.723), but the control group exhibited a higher cadence during the constant vs varying condition (p < 0.001).

Testing for implicit and explicit retention effects on cadence following overground testing revealed a main effect of cue (p < 0.001, ηp2 = 0.923), and an interaction effect of cue × metronome × group (p = 0.032, ηp2 = 0.333) (Figure 5a). When split by group, there was no interaction effect of cue × metronome on cadence for the older adult controls (p = 0.511), however, we observed a cue × metronome interaction for PwPD (p = 0.002, ηp2 = 0.792) such that the cadence during implicit retention testing during the varying condition was significantly higher than that of the constant condition (p = 0.037). There was no difference in cadence between the isochronous and varying metronomes during baseline and explicit retention testing for PwPD (p = 0.867 and p = 0.160, respectively). Across both groups, cadence was higher in the implicit and explicit retention trials than baseline, and higher in the explicit as compared to implicit retention trial (all p < 0.001). There was a main effect of cue only for gait speed (p < 0.001, ηp2 = 0.865), such that implicit and explicit retention testing yielded greater gait speed than baseline walking, and explicit retention testing yielded greater gait speed than implicit retention testing (all p ≤ 0.002) (Figure 5b).

Fig. 5.

Fig. 5

Open in a new tab

Implicit and explicit retention effects on (A) cadence and (B) gait speed after timing steps to the isochronous (open shape) and subtly varying (closed shape) metronomes. Cadence and gait speed are represented as the percentage of comfortable overground cadence and gait speed, respectively. Baseline refers to the participants overground cadence and gait speed prior to introduction of the metronome during overground training. These may differ from the initial comfortable overground cadence and gait speed, as all participants first underwent the treadmill condition (See Figure 1 for details). Each participant’s average cadence and step length are represented by a data point on the figure

Discussion

Our hypothesis that older adults and PwPD would change their cadence based on subtle variations in metronome cues, without conscious recognition of the change, both on a treadmill and overground was only partially supported by the data. Neither group appeared to change cadence while walking on the treadmill with the varying metronome. However, despite 9 out of 10 participants in each group being unaware the metronome was changing frequency, both groups adjusted their overground cadence in response to these changes. Thus, our results suggest that incorporating error-based recalibration into RAC-guided overground gait training may be feasible for both older adults and PwPD, even though recalibration did not occur in all participants during treadmill walking with the varying metronome.

The observed persistence in cadence with the subtly varying metronome on the treadmill may be explained by the relationship between step length, cadence, and gait speed. When speed is held constant, as it was on the treadmill, reducing cadence necessitates an increase in step length. Thus, the consistent cadence observed with the subtly varying metronome may be the result of a broader challenge associated with the need to increase step length to accommodate for the slower cadence on the fixed speed treadmill (Chawla et al. 2020). Some participants in both groups demonstrated difficulty reaching 85% of their comfortable overground cadence while on the treadmill, as illustrated in Figure 2. Reduced step lengths are already ubiquitous in individuals with PD and otherwise unimpaired older adults (Judge et al. 1996; Mak 2013; Galna et al. 2015; Shearin et al. 2021; Welzel et al. 2021), making the need to take longer steps particularly challenging. Given the participants’ difficulty entraining to the 85% metronome cue, it is not surprising that they had trouble following the subtle variations of that cue.

Furthermore, PwPD demonstrate difficulty modulating step length (Morris et al. 1994; Bayle et al. 2016), which may have impacted our participants’ ability to synchronize with the metronome’s subtle changes in frequency. Although studies indicate that young, unimpaired adults can adapt their cadence to a varying metronome while on a treadmill (Forner-Cordero et al. 2019; Duppen et al. 2023), a combination of age- and neuropathology-related reductions in step length and impaired ability to modulate step length may have posed too great a challenge for some of the participants in both groups.

Another factor contributing to participants’ difficulty in synchronizing with both the isochronous and varying metronomes on the treadmill could be the lack of optic flow. During treadmill walking, visual information indicates that the individual is nearly stationary, while proprioceptive input from the lower extremities signal movement. This mismatch between visual and proprioceptive cues can both impair performance and hinder the transfer of learned gait behaviors.(Torres-Oviedo and Bastian 2010) Therefore, the nature of treadmill walking itself may have made it more challenging to match the metronome frequencies accurately for our older adults and PwPD, who tend to rely more heavily on vision for motor control during walking.(Berard et al. 2012; Yakubovich et al. 2020; Tran et al. 2023)

Conversely, both groups altered their cadence, without conscious recognition of the change, while walking overground to a subtly varying metronome. Coincidently, participants appeared to have greater success at matching the 15% increase in target cadence as compared to the 15% decrease in target cadence on the treadmill. Walking overground is less constrained than walking on a treadmill that is set to a fixed speed. For instance, when walking overground participants have the options of synchronizing with the metronome frequency by adopting shorter steps, maintaining similar length steps as their baseline, or even increasing step length, as their gait speed is not fixed. In fact, prior work suggests that RAC overground does not result in significant changes to step length (Hoppe et al. 2020). This ability to disentangle the spatial and temporal control features of gait, and the lack of requirement to increase step lengths, likely accounts for our participants’ ability to entrain to the subtly varying metronome while walking overground, despite showing less ability to do so on the treadmill.

Our second hypothesis, that a slowly varying metronome would elicit greater implicit retention of altered gait behaviors compared to the isochronous metronome, was also only partially supported by the results. Neither group substantially altered their cadence to match the subtly varying slow-frequency metronome cue during treadmill walking (i.e., they did not demonstrate error-based recalibration), which likely explains the absence of differences in implicit retention effects between the isochronous and varying metronomes. However, during overground walking, only PwPD showed an increase in implicit retention effects on cadence after walking with the varying metronome compared to the isochronous metronome. While this result does not support our hypothesis that both groups would demonstrate increased implicit retention effects after walking to the subtly varying versus isochronous metronome, it does highlight the ability of PwPD to employ implicit motor learning approaches during gait training overground.

Our results support the idea that PwPD retain the ability to adopt gait behaviors via implicit motor learning approaches during overground walking using error-based recalibration. Historically, people with neurologic conditions affecting the striatum, such as those with PD, were considered unable to learn implicitly (Vandenbossche et al. 2013; Beigi et al. 2016). However, current motor learning theory suggests there are two distinct processes involved in implicit motor learning, with the initial phase (i.e., rapid encoding) dependent on the hippocampus, and the later phase (i.e., proceduralization) requiring the striatum (Henke 2010; Nemeth et al. 2013; Gamble et al. 2014). This distinction is especially important for PwPD, as the hippocampus is less affected by disease processes. In fact, during finger tapping tasks, individuals with PD are able to learn implicitly at a similar rate and accuracy as unimpaired older adult controls when the task is changed frequently enough for participants to remain within the initial phase of implicit learning (Gamble et al. 2014). Error-based recalibration involves continuous adjustment of movements, fully engaging the rapid encoding phase of implicit learning. Because this recalibration, driven by sensory prediction errors, appears to occur primarily in the cerebellum (Tseng et al. 2007) and relies on the hippocampus for rapid encoding, it may offer a viable alternative to conventional strategic and reinforcement learning approaches for PwPD.

Moreover, implicit-weighted motor learning mechanisms, such as error-based recalibration, are linked to enhanced movement automaticity (Kal et al. 2018), improved retention of learned behaviors (Lee and Vakoch 1996; Kim et al. 2015), and are considered less cognitively demanding than more explicit-weighted methods (Mazzoni and Krakauer 2006). Given that PwPD often struggle with movement automaticity and face challenges with cognitive-motor dual tasks (O’Shea et al. 2002; Wu and Hallett 2005; Raffegeau et al. 2019), incorporating error-based recalibration to infuse implicit-weighted motor learning into gait training has the potential to greatly benefit this population.

While results were promising for the ability to use implicit learning approaches to alter gait behaviors for PwPD during overground walking, we acknowledge several limitations to this study. First, the effect sizes of changes in overground cadence during the varying condition were quite small. This is not surprising, however, as we altered the intended cadence by only one percent of comfortable overground cadence every 20 seconds. Therefore, the maximum cued change was only a 6% change from their comfortable cadence. Importantly, despite these small effect sizes between epochs, effect sizes during implicit and explicit retention testing were quite large.

Another limitation of this study is that our cohort with PD demonstrated only mild to moderate impairment according to self-selected comfortable gait speed, step length, cadence, Hoehn & Yahr staging, and UPDRS motor scores, making it difficult to generalize these results to individuals with more severe impairment. We also did not find a difference in gait speed or step length between our control group and PD group, which is inconsistent with a multitude of studies reporting smaller step lengths and slower gait speed for PwPD compared to people without PD (Morris et al. 1994; Galna et al. 2015). This lack of difference between our two groups is likely due to our PwPD being highly active. For instance, nearly all members of the PwPD group actively engaged in various physical activities such as tai chi, boxing, PWR! classes, and pickleball. Moderate physical activity is known to be neuroprotective (Palasz et al. 2019), and has been associated with improved gait parameters for PwPD (Zhen et al. 2022). Therefore, it is important to note that our cohort may not fully represent a broader population of PwPD. Another possible reason for the lack of difference between groups is our inclusion/exclusion criteria. We included PwPD in Hoehn & Yahr stages I-III without screening for gait-related hypokinesia or bradykinesia. Future studies might consider including these criteria, as rehabilitation specialists typically focus gait training on individuals experiencing these specific gait impairments.

Moreover, despite no known neurologic impairment in our control group, and mild neurologic impairment in our PwPD group, neither group was able to consistently match the 85% metronome on the treadmill or appropriately vary their treadmill cadence based on the subtly varying metronome. However, both groups did demonstrate the ability to reduce cadence and increase step length during both treadmill conditions. This could be indicative of the 85% metronome cue being too large a challenge for our groups to overcome.

Lastly, a limitation of this work, and others, is that the use of implicit motor learning approaches in PD only assess exposure (i.e., single session) rather than true longitudinal, long-term training. With a distinction between rapid encoding and proceduralization within implicit learning (Henke 2010; Nemeth et al. 2013; Gamble et al. 2014), it will be important to determine whether difficulty with proceduralization will impact longer-term retention of gait changes. Future research will be needed to determine if longer-term motor learning can be achieved through error-based recalibration in PwPD.

Funding:

This research was funded in part by a Promotion of Doctoral Studies (PODS) I Scholarship from the Foundation for Physical Therapy Research [944183]. The project described was also supported by the National Center for Advancing Translational Sciences, National Institutes of Health, through Grant Award Number TL1TR002491, and by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, through Grant Award Number R21HD111833. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Footnotes

Conflict of Interest: The authors declare that they have no conflict of interest.

Data Availability:

The data that support the findings of this study are openly available in UNC Dataverse at https://doi.org/10.15139/S3/U7CWWN.

References

  1. Ashoori A, Eagleman DM, Jankovic J (2015) Effects of Auditory Rhythm and Music on Gait Disturbances in Parkinson’s Disease. Front Neurol 6:234 doi: 10.3389/fneur.2015.00234 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Azulay JP, Mesure S, Amblard B, Blin O, Sangla I, Pouget J (1999) Visual control of locomotion in Parkinson’s disease. Brain 122 ( Pt 1):111–120 doi: 10.1093/brain/122.1.111 [DOI] [PubMed] [Google Scholar]
  3. Bayle N, Patel AS, Crisan D, et al. (2016) Contribution of Step Length to Increase Walking and Turning Speed as a Marker of Parkinson’s Disease Progression. PLoS One 11:e0152469 doi: 10.1371/journal.pone.0152469 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Beigi M, Wilkinson L, Gobet F, Parton A, Jahanshahi M (2016) Levodopa medication improves incidental sequence learning in Parkinson’s disease. Neuropsychologia 93:53–60 doi: 10.1016/j.neuropsychologia.2016.09.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Berard J, Fung J, Lamontagne A (2012) Impact of aging on visual reweighting during locomotion. Clin Neurophysiol 123:1422–1428 doi: 10.1016/j.clinph.2011.11.081 [DOI] [PubMed] [Google Scholar]
  6. Chawla G, Hoppe M, Browner N, Lewek MD (2020) Individuals With Parkinson’s Disease Retain Spatiotemporal Gait Control With Music and Metronome Cues. Motor Control:1–11 doi: 10.1123/mc.2020-0038 [DOI] [PubMed] [Google Scholar]
  7. Chen HY, Wing AM, Pratt D (2006) The synchronisation of lower limb responses with a variable metronome: the effect of biomechanical constraints on timing. Gait Posture 23:307–314 doi: 10.1016/j.gaitpost.2005.04.001 [DOI] [PubMed] [Google Scholar]
  8. Del Olmo MF, Cheeran B, Koch G, Rothwell JC (2007) Role of the cerebellum in externally paced rhythmic finger movements. J Neurophysiol 98:145–152 doi: 10.1152/jn.01088.2006 [DOI] [PubMed] [Google Scholar]
  9. Duppen CP, Wrona H, Dayan E, Lewek MD (2023) Evidence of Implicit and Explicit Motor Learning during Gait Training with Distorted Rhythmic Auditory Cues. J Mot Behav:1–10 doi: 10.1080/00222895.2023.2231874 [DOI] [PubMed] [Google Scholar]
  10. Forner-Cordero A, Pinho JP, Umemura G, Lourenco JC, Mezencio B, Itiki C, Krebs HI (2019) Effects of supraspinal feedback on human gait: rhythmic auditory distortion. J Neuroeng Rehabil 16:159 doi: 10.1186/s12984-019-0632-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. French MA, Cohen ML, Pohlig RT, Reisman DS (2021) Fluid Cognitive Abilities Are Important for Learning and Retention of a New, Explicitly Learned Walking Pattern in Individuals After Stroke. Neurorehabil Neural Repair 35:419–430 doi: 10.1177/15459683211001025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fujioka T, Trainor LJ, Large EW, Ross B (2012) Internalized timing of isochronous sounds is represented in neuromagnetic beta oscillations. J Neurosci 32:1791–1802 doi: 10.1523/jneurosci.4107-11.2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Galna B, Lord S, Burn DJ, Rochester L (2015) Progression of gait dysfunction in incident Parkinson’s disease: impact of medication and phenotype. Mov Disord 30:359–367 doi: 10.1002/mds.26110 [DOI] [PubMed] [Google Scholar]
  14. Gamble KR, Cummings TJ Jr, Lo SE, Ghosh PT, Howard JH Jr, Howard DV (2014) Implicit sequence learning in people with Parkinson’s disease. Front Hum Neurosci 8:563 doi: 10.3389/fnhum.2014.00563 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ghai S, Ghai I, Schmitz G, Effenberg AO (2018) Effect of rhythmic auditory cueing on parkinsonian gait: A systematic review and meta-analysis. Sci Rep 8:506 doi: 10.1038/s41598-017-16232-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Henke K (2010) A model for memory systems based on processing modes rather than consciousness. Nat Rev Neurosci 11:523–532 doi: 10.1038/nrn2850 [DOI] [PubMed] [Google Scholar]
  17. Hoppe M, Chawla G, Browner N, Lewek MD (2020) The effects of metronome frequency differentially affects gait on a treadmill and overground in people with Parkinson disease. Gait Posture 79:41–45 doi: 10.1016/j.gaitpost.2020.04.003 [DOI] [PubMed] [Google Scholar]
  18. Hsu JL, Fan YC, Huang YL, Wang J, Chen WH, Chiu HC, Bai CH (2015) Improved predictive ability of the Montreal Cognitive Assessment for diagnosing dementia in a community-based study. Alzheimers Res Ther 7:69 doi: 10.1186/s13195-015-0156-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Jablonowski J, Taesler P, Fu Q, Rose M (2018) Implicit acoustic sequence learning recruits the hippocampus. PLoS One 13:e0209590 doi: 10.1371/journal.pone.0209590 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Judge JO, Davis RB 3rd, Ounpuu S (1996) Step length reductions in advanced age: the role of ankle and hip kinetics. J Gerontol A Biol Sci Med Sci 51:M303–312 doi: 10.1093/gerona/51a.6.m303 [DOI] [PubMed] [Google Scholar]
  21. Kal E, Prosee R, Winters M, van der Kamp J (2018) Does implicit motor learning lead to greater automatization of motor skills compared to explicit motor learning? A systematic review. PLoS One 13:e0203591 doi: 10.1371/journal.pone.0203591 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kim SJ, Ogilvie M, Shimabukuro N, Stewart T, Shin JH (2015) Effects of Visual Feedback Distortion on Gait Adaptation: Comparison of Implicit Visual Distortion Versus Conscious Modulation on Retention of Motor Learning. IEEE Trans Biomed Eng 62:2244–2250 doi: 10.1109/tbme.2015.2420851 [DOI] [PubMed] [Google Scholar]
  23. King LA, Priest KC, Salarian A, Pierce D, Horak FB (2012) Comparing the Mini-BESTest with the Berg Balance Scale to Evaluate Balance Disorders in Parkinson’s Disease. Parkinsons Dis 2012:375419 doi: 10.1155/2012/375419 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lee YS, Vakoch DA (1996) Transfer and retention of implicit and explicit learning. Br J Psychol 87 ( Pt 4):637–651 doi: 10.1111/j.2044-8295.1996.tb02613.x [DOI] [PubMed] [Google Scholar]
  25. Leow LA, Marinovic W, de Rugy A, Carroll TJ (2018) Task errors contribute to implicit aftereffects in sensorimotor adaptation. Eur J Neurosci 48:3397–3409 doi: 10.1111/ejn.14213 [DOI] [PubMed] [Google Scholar]
  26. Long AW, Roemmich RT, Bastian AJ (2016) Blocking trial-by-trial error correction does not interfere with motor learning in human walking. J Neurophysiol 115:2341–2348 doi: 10.1152/jn.00941.2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mak MK (2013) Reduced step length, not step length variability is central to gait hypokinesia in people with Parkinson’s disease. Clin Neurol Neurosurg 115:587–590 doi: 10.1016/j.clineuro.2012.07.014 [DOI] [PubMed] [Google Scholar]
  28. Mazzoni P, Krakauer JW (2006) An implicit plan overrides an explicit strategy during visuomotor adaptation. J Neurosci 26:3642–3645 doi: 10.1523/jneurosci.5317-05.2006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Morris ME, Iansek R, Matyas TA, Summers JJ (1994) The pathogenesis of gait hypokinesia in Parkinson’s disease. Brain 117 ( Pt 5):1169–1181 doi: 10.1093/brain/117.5.1169 [DOI] [PubMed] [Google Scholar]
  30. Nemeth D, Janacsek K, Kiraly K, et al. (2013) Probabilistic sequence learning in mild cognitive impairment. Front Hum Neurosci 7:318 doi: 10.3389/fnhum.2013.00318 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Nieuwboer A, Kwakkel G, Rochester L, et al. (2007) Cueing training in the home improves gait-related mobility in Parkinson’s disease: the RESCUE trial. J Neurol Neurosurg Psychiatry 78:134–140 doi: 10.1136/jnnp.200X.097923 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. O’Shea S, Morris ME, Iansek R (2002) Dual task interference during gait in people with Parkinson disease: effects of motor versus cognitive secondary tasks. Phys Ther 82:888–897 [PubMed] [Google Scholar]
  33. Owens-Walton C, Jakabek D, Li X, et al. (2018) Striatal changes in Parkinson disease: An investigation of morphology, functional connectivity and their relationship to clinical symptoms. Psychiatry Res Neuroimaging 275:5–13 doi: 10.1016/j.pscychresns.2018.03.004 [DOI] [PubMed] [Google Scholar]
  34. Palasz E, Niewiadomski W, Gasiorowska A, Wysocka A, Stepniewska A, Niewiadomska G (2019) Exercise-Induced Neuroprotection and Recovery of Motor Function in Animal Models of Parkinson’s Disease. Front Neurol 10:1143 doi: 10.3389/fneur.2019.01143 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Puyjarinet F, Begel V, Geny C, et al. (2019) Heightened orofacial, manual, and gait variability in Parkinson’s disease results from a general rhythmic impairment. NPJ Parkinsons Dis 5:19 doi: 10.1038/s41531-019-0092-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Raffegeau TE, Krehbiel LM, Kang N, Thijs FJ, Altmann LJP, Cauraugh JH, Hass CJ (2019) A meta-analysis: Parkinson’s disease and dual-task walking. Parkinsonism Relat Disord 62:28–35 doi: 10.1016/j.parkreldis.2018.12.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Repp BH (2000) Compensation for subliminal timing perturbations in perceptual-motor synchronization. Psychol Res 63:106–128 doi: 10.1007/pl00008170 [DOI] [PubMed] [Google Scholar]
  38. Richardson JT (2011) Eta squared and partial eta squared as measures of effect size in educational research. Educational research review 6:135–147 [Google Scholar]
  39. Shearin SM, Medley A, Trudelle-Jackson E, Swank C, Querry R (2021) Plantarflexor strength, gait speed, and step length change in individuals with Parkinson’s disease. Int J Rehabil Res 44:82–87 doi: 10.1097/MRR.0000000000000439 [DOI] [PubMed] [Google Scholar]
  40. Sherron MA, Stevenson SA, Browner NM, Lewek MD (2020) Targeted Rhythmic Auditory Cueing During Treadmill and Overground Gait for Individuals With Parkinson Disease: A Case Series. J Neurol Phys Ther doi: 10.1097/npt.0000000000000315 [DOI] [PubMed] [Google Scholar]
  41. Taylor JA, Krakauer JW, Ivry RB (2014) Explicit and implicit contributions to learning in a sensorimotor adaptation task. J Neurosci 34:3023–3032 doi: 10.1523/jneurosci.3619-13.2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Thaut MH, Rice RR, Braun Janzen T, Hurt-Thaut CP, McIntosh GC (2019) Rhythmic auditory stimulation for reduction of falls in Parkinson’s disease: a randomized controlled study. Clin Rehabil 33:34–43 doi: 10.1177/0269215518788615 [DOI] [PubMed] [Google Scholar]
  43. Torres-Oviedo G, Bastian AJ (2010) Seeing is believing: effects of visual contextual cues on learning and transfer of locomotor adaptation. J Neurosci 30:17015–17022 doi: 10.1523/JNEUROSCI.4205-10.2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Tran S, Brooke C, Kim YJ, et al. (2023) Visual and vestibular integration in Parkinson’s disease while walking. Parkinsonism Relat Disord 116:105886 doi: 10.1016/j.parkreldis.2023.105886 [DOI] [PubMed] [Google Scholar]
  45. Tseng YW, Diedrichsen J, Krakauer JW, Shadmehr R, Bastian AJ (2007) Sensory prediction errors drive cerebellum-dependent adaptation of reaching. J Neurophysiol 98:54–62 doi: 10.1152/jn.00266.2007 [DOI] [PubMed] [Google Scholar]
  46. Vandenbossche J, Deroost N, Soetens E, et al. (2013) Impaired implicit sequence learning in Parkinson’s disease patients with freezing of gait. Neuropsychology 27:28–36 doi: 10.1037/a0031278 [DOI] [PubMed] [Google Scholar]
  47. Welzel J, Wendtland D, Warmerdam E, et al. (2021) Step Length Is a Promising Progression Marker in Parkinson’s Disease. Sensors (Basel) 21 doi: 10.3390/s21072292 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Willems AM, Nieuwboer A, Chavret F, et al. (2006) The use of rhythmic auditory cues to influence gait in patients with Parkinson’s disease, the differential effect for freezers and non-freezers, an explorative study. Disabil Rehabil 28:721–728 doi: 10.1080/09638280500386569 [DOI] [PubMed] [Google Scholar]
  49. Wu T, Hallett M (2005) A functional MRI study of automatic movements in patients with Parkinson’s disease. Brain 128:2250–2259 doi: 10.1093/brain/awh569 [DOI] [PubMed] [Google Scholar]
  50. Yakubovich S, Israeli-Korn S, Halperin O, Yahalom G, Hassin-Baer S, Zaidel A (2020) Visual self-motion cues are impaired yet overweighted during visual-vestibular integration in Parkinson’s disease. Brain Commun 2:fcaa035 doi: 10.1093/braincomms/fcaa035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zhen K, Zhang S, Tao X, Li G, Lv Y, Yu L (2022) A systematic review and meta-analysis on effects of aerobic exercise in people with Parkinson’s disease. NPJ Parkinsons Dis 8:146 doi: 10.1038/s41531-022-00418-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are openly available in UNC Dataverse at https://doi.org/10.15139/S3/U7CWWN.

RESOURCES