Factors associated with weight-shift amplitude at Baseline
To our knowledge, this is the first instance where sufficient data has been generated to afford statistical comparisons of APA characteristics between: (1) successful and unsuccessful attempts to step from a freeze (within-subject) and ii) attempts to initiate gait from both a freeze (FoG-F) and a voluntary stop (FoG-NF). We attribute this success in inducing FoG in our laboratory setting to the use of the SIP task in conjunction with the VR environments designed specifically to represent scenarios commonly associated with exacerbated FoG events (e.g., narrow doorways). As such, while we acknowledge limitations associated with using VR head-mounted displays and the necessary contrived laboratory environment/task, we also suggest that our current data may, in comparison to extant literature, better reflect behaviour of patients attempting to initiate gait in daily life.
Previous work has documented impoverished APAs in PD with FoG when initiating gait in the absence of any observed freeze [
13,
26], leading to suggestions that FoG may be a consequence of attenuated/defective APA scaling [
8]. Our findings provide novel and additional support for this notion, as ML-WS amplitudes were significantly attenuated during attempted steps that were unsuccessful (Fig.
3).
Schlenstedt et al. [
14] compared the magnitude of ML APAs between groups of PD patients that were categorised according to presence/absence of self-reported FoG pathology. They reported no significant between-group differences at baseline during successfully initiated steps; findings that conflict with ‘traditional’ conceptualisations described above. Their study also provided data relating to nine attempted steps where start hesitations were observed and determined that APA amplitude was relatively large compared to step initiations where FoG was absent. The authors interpreted these trends as representing a potential compensatory mechanism aimed at avoiding destabilising accelerations in the centre of mass. These interpretations were supported by positive associations observed between self-reported FoG and APA amplitude [
14] and velocity [
52]. Our data indicate that no such relationship exists, at least not concerning our measure of ML-WS (Fig.
4 C&E); a finding that is consistent across participant groups (i.e., it is not contingent on whether individuals are attempting to step from a freeze or voluntary stop). However, despite this apparent discrepancy, we argue for a similar interpretation and conclusion. Our findings from the FoG-F group at Baseline show that smaller ML-WS amplitudes are associated with unsuccessful attempts to step. Therefore, occasions where participants increased ML-WS (and successfully initiated steps from a freeze) might indeed represent an effective compensatory strategy to overcome FoG, as suggested by Schlenstedt et al. [
13]. The question remains whether users must be cognisant of processes involved in the production of larger weight-shifts in order to attain improvements in motor performance.
We must be cautious when contextualising the evident relationship between increased ML-WS and the improved rate of successful gait initiations (Fig.
6). Specifically, it is unclear if the hypokinetic ML-WS described during unsuccessful attempts to initiate walking (Fig.
3) should be considered as a mechanism primarily responsible for resultant failed attempts, or as one of several integrated behaviours that are symptomatic of more fundamental defective mechanisms (i.e., an artefact of the failed step). Similarly, the production of larger ML-WS during experimental conditions may not represent the most important factor driving improvements in step initiation. For example, benefits may have been derived from internally generated rhythmic cues corresponding to the components of the APA (Fig.
1) or relative increases in ML-WS duration, although the latter is a likely consequence of increased ML-WS amplitude. We suggest that the observed improvements to gait initiation may represent a combination of the above, driven largely by efficiencies associated with allocating attention specifically towards the production of an APA.
Several studies have contextualised abnormal preparatory movements ahead of attempted gait initiation as representing failed attempts to generate [
53] and/or inhibit [
20,
54] APA-specific motor programmes. Carlsen et al. [
55] suggested that when a motor task (such as gait initiation) is known in advance, cortical demands associated with movement planning and execution can be reduced via the discharge and subsequent inhibition of any relevant motor program to a so-called ‘holding area’. This stored motor program is then subject to self-generated or externally triggered release. Dividing attention may therefore compromise this tonic inhibition, thus releasing APAs that are spatially and/or temporally defective. The notion that defective inhibition of APAs may influence freezing is supported by reports that dual-task paradigms can exacerbate FoG symptoms [
56]. As such, external sensory cueing aimed at facilitating gait initiation may be problematic if attending to the cue (and processing the information presented) carries attentional demands that are not directly associated with the most relevant aspects of the ‘primed’ motor programmes.
Many studies have attempted to utilise sensory cues to facilitate gait initiation in people with FoG. However, such strategies have typically specified information regarding the ultimate goal of initiating movement in the stepping limb or completing the initiated step in a given manner (e.g., placing the foot on/over a visual target [
5], or imitating the sound [
45] or mental image of a stepping action [
57,
58]. These attempts to focus attention towards step-related movement goals are reflected in 26/35 of our participants’ self-reporting attempts to consciously initiate foot movement at Baseline (e.g., “pick your foot up”, “put your heel down first”). Therefore, the transition between Baseline to Verbal trials was not characterised by the adoption of conscious strategies per se, but rather a change in the
allocation of attention between aspects of motor performance (i.e., from the production of a resultant step to an integrated weight-shift and step action) that was sufficient to induce observed improvements (Fig.
6).
It is notable that the relative absence of attentional focus on the resultant successful forward step (in Verbal and Analogy conditions, see Table
2) is indicative of this action maintaining a degree of automaticity; perhaps being triggered by the preceding augmented APA as part of a ‘domino effect’. In this sense, attentional focus on intended foot (rather than APA) movement reported at Baseline might represent attempts to affect the final component of this chain-reaction, resulting in a higher proportion of failed attempts (Fig.
6).
We propose that the newly adopted attentional focus specifically towards APA production may have avoided the need for participants to generate, store and inhibit an independent APA motor programme that would likely be susceptible to problems associated with premature and ineffective release. We propose that the generation of suitably scaled APAs were initiated through, and coupled to, specific pre-potent conscious strategies. This process would, in theory, reduce the burden on deficient ‘automatic’ processes [
24,
59,
60] and improve cognitive processing efficiency. More specifically, the above suggestion may constitute a reduction in demands placed on any ‘supervisory system’ [
61] otherwise employed to create tonic inhibition on APA motor programmes.
In the current study, we predicted that the use of analogies as a strategy to increase ML-WS would deliver greater benefits, by virtue of potential reductions in attentional demands, compared to the Verbal condition. We observed no evidence of this. Despite FoG-F ML-WS being relatively smaller in Baseline trials (Fig.
5), participants (stepping from a freeze) were able to achieve equivalent ML-WS amplitudes, and successfully initiate forward steps (approx. 100%), during both Verbal and Analogy conditions (Fig.
6). Indeed, the magnitude of improvement in the ratio of successful/unsuccessful steps makes it difficult to make direct comparisons between Verbal and Analogy conditions. Typically, the virtues of implicit approaches to motor learning are evident when comparative verbal processes are complex and vulnerable to breaking down under anxiety. As such, it is possible that verbal strategies used in the current study to initiate gait were sufficiently simple to avoid such issues. After all, descriptions of the APA during training were accompanied by uncomplicated verbal cues that participants were encouraged to use during attempted step initiation (e.g., counting “One, Two, Three” to mark each phase of the APA). Given the weight of evidence from other contexts [
30], we maintain that motor performance would show deficiencies if more complex verbal processes were evident, particularly during anxiogenic tasks.
In contrast to the simple verbal cues described above, one might consider the APA-relevant analogies as being relatively complex, and therefore potentially imparting greater demands on cognitive resources relative to the requirements of the task. It is possible that any negative consequences of these increased demands may have been mitigated by the benefits of having internally generated a rich dynamic template to which participants could adhere their movement. In the current study, parity in outcomes observed between Verbal and Analogy conditions suggests that benefits to ML-WS and successful step initiation do not exclusively rely on the explicit conscious control of APA mechanics. This demonstrates clear potential for researchers and clinicians to develop creative solutions for specifying spatial and temporal APA characteristics that are tailored to individual patients’ preferences and requirements. To this end, it should be noted that, in the specific context of APA and step initiation, external sensory cueing strategies carry the following limitation.
Coupling one’s movement to an external source of spatiotemporal information is typically most successful during cyclical and continuous tasks (e.g., finger tapping or walking) as it allows the performer to use estimates of previous errors to adjust and improve subsequent actions. An APA and subsequent step initiation represents a discrete and complex task that cannot be readily adapted and made continuous (and therefore more amenable to external cueing). Therefore, aside from potential logistical issues associated with organising external devices to deliver sensory cues, we propose that internally-generated verbal or imagery cues (such as those evaluated here), represent the most practical solution for people attempting to overcome FoG in daily life. The significant limitation to this suggestion relates to cognitive deficits that have been (albeit inconsistently) associated with FoG (see Gilat and Colleagues [
4] for review).
The current study did not attempt to document freezing ‘sub-types’ or categorise participants accordingly [
62]. As such, we cannot draw any conclusions regarding the utility of weight-shifting between sub-types or even infer if the results observed are representative of all sub-types. Further work should explore this issue in addition to the potential impacts of specific deficits in cognitive and executive function on the use of the strategies described.
Given the clear improvements observed in the rate of successful gait initiation during both Verbal and Analogy conditions, we suggest that these strategies might represent an effective non-pharmacological option for the practical management of FoG that could be trained in conventional rehabilitation settings. However, given the contrived nature of our stepping tasks and VR environment (including the limitation that participants were required to wear a physical headset), it is important to further evaluate the efficacy and safety of the approaches described in community settings. The current study was also limited by our singular focus on forward walking, meaning that the effects observed cannot be assumed to translate to other variants of attempted gait initiation, such as attempts to turn and change walking direction. Further work is necessary to develop similar strategies to those described here in an attempt to facilitate attempted turns during FoG. We also suggest that future work should aim to record more detailed descriptions of thought processes evident when people attempt to use weight-shifting strategies to overcome FoG in daily life and evaluate possible associations with movement outcomes.