Real-time augmented feedback for gait training: are gait responses affected by the choice of target parameter?

Augmented visual feedback (AVF) is a promising approach for gait rehabilitation after stroke. However, we still lack crucial knowledge about how to most efficiently use it.

How does the selection of the gait parameter targeted by the AVF signal influence the global motor response (i.e., overall gait pattern)?

24 healthy young participants (mean age 25.3 $\pm$ 5.3 years old) performed one session of treadmill walking while receiving real-time AVF driving them towards gait asymmetry. AVF was given during 3 × 10 mins on three different gait parameters: stance time (ST), ankle plantarflexion at toe-off (APL) and push-off force (POF). We analyzed gait responses throughout, with a focus on parameter-specific (local) changes in symmetry ratio, and global gait pattern changes, quantified by correlation coefficient and Gait Deviation Index (GDI).

When using ST and POF feedback targets, participants successfully modified their local asymmetry by an average of 10%. Correlation analysis (Spearman) indicated that the modulated gait propagated across parameters, with a fair correlation $\left(\left|r\right|=0.3-0.5\right)$ between ST and APL, POF or vertical ground reaction force and between POF and ST, swing time, step length, step height or vertical ground reaction force. However, global gait pattern was only negatively influenced by ST feedback (GDI -7.9 points with $p\le 0.01$ ). Conversely, APL did not lead to significant local symmetry modulation.

Our results show that the efficacy of AVF is dependent on the selected target parameter. This choice also seems to affect how local symmetry changes affect global motion patterns. This work is a first step towards a more comprehensive understanding of the direct and indirect impact of AVF on gait response, which is crucial before using AVF for clinical applications.