Effects of augmented reality cueing strategies on freezing of gait: The ELIMINATE FoG trial

Abstract

Background

Freezing of gait (FoG) is a treatment-resistant symptom of Parkinson disease (PD). Augmented reality (AR) cues have been investigated as a therapy for FoG, with inconclusive results from a limited array of AR constructs.

Objectives

Compare four modalities of a novel AR cue to physical and no-cue controls.

Methods

Presence of FoG in PD was required; exclusion criteria included dementia, severe vision loss, and significant gait-disrupting comorbidities. Participants completed six walking tasks, featuring different cueing conditions in a crossover fashion, in a holographic hallway displayed by an AR headset. A conventional physical cue was presented first, followed by other conditions in randomized order (hand-controlled AR cue, observer-controlled AR cue, eye-controlled AR cue, constant AR cue, no-cue control). Primary outcomes were FoG duration and incidence, manually annotated. Secondary outcomes included survey questions and gait parameters derived from IMUs.

Results

Thirty-six participants completed testing. The observer-controlled AR cue produced lower FoG duration than the no-cue, physical, and hand-controlled AR cue conditions (N = 36, p ≤ 0.006, Wilcoxon effect size (WES) ≥ 0.46). The constant cue reduced FoG incidence compared to all other conditions (N = 36, p ≤ 0.016, WES ≥ 0.40). Participants’ preferred AR cues decreased FoG duration (N = 28, p ≤ 0.004, WES ≥ 0.48) and incidence (N = 28, p ≤ 0.022, WES ≥ 0.38) compared to controls. Differences in kinematic outcomes were negligible. Survey results indicated receptiveness toward AR cueing, with diversity in preferred cue activation modalities. No significant adverse events occurred.

Conclusions

AR cueing decreased FoG incidence and duration compared to controls. Efficacy of discrete cueing modalities likely depends on user intrinsic factors, such as preference.