Table 1 Summary of the perturbation paradigms and results of the included studies

Bhatt et al. [55]

Single session: n = 25, 13♀, 73.42 ± 5.42y. Dual session: n = 13, 7♀, 70.13 ± 4.75y. All community dwelling, healthy older adults

Slip (low friction moveable platform; slid up to 90 cm forward at foot contact).

Single session: 24 slips in 37 gait trials. Dual session: as above, plus 1 slip trial 3mo later.

Stability, loss of balance & hip height during slip 1 & 24, & the single slip at +6mo. % of falls following the lab perturbations.

Improvement in the observed parameters from slip 1 to 24. Retention in all parameters at +6mo, greater for the dual group.

Bierbaum et al. [64]

13♂, 67.4 ± 3.4y. Community dwelling, healthy.

Surface change perturbation (17 cm thick foam with an average of 10 cm deformation).

19 trials with the 2nd, 8th & 19th as a hard surface, & the rest soft surface.

MoS, BoS & X_CoM at touchdown of perturbed & recovery steps.

Improvement across trials in the outcome parameters.

Bierbaum et al. [65]

14♂, 67.3 ± 4.2y. Community dwelling, healthy.

Surface change perturbation (17 cm thick foam with an average of 10 cm deformation).

28 trials: 23 with hard surface, 5 with soft surface.

MoS, BoS & X_CoM at touchdown of the perturbed & recovery steps.

Improved MoS of the recovery step for the 4th & 5th perturbations compared to the 1st.

Lurie et al. [56]

n = 26, 13♀, 81 ± 6.53y. Healthy older adults referred for gait and balance training.

Anterior/posterior treadmill accelerations of progressive magnitude (scale of 1–5, exact values not reported).

5.84 sessions of 44.25mins (means). Therapist determined perturbation type (stance or gait), magnitude & number.

Mean perturbation magnitude successfully negotiated per session. Retrospective falls data 3mo preceding & for 3mo after the intervention.

Improved mean trip magnitude from 2.44 to 3.44. Non-significant difference in subjects experiencing falls (19% vs. 33%) compared to controls.

Pai et al. [62]

n = 38, 19♀, 71 ± 5y. Community dwelling, healthy.

Slip (low friction moveable platform; slid up to 90 cm forward at foot contact).

24 slips in 37 gait trials.

Stability, loss of balance & hip height 300 ms after perturbation onset. % of falls following the lab perturbations.

Reduction in falls & backward losses of balance across trials. Improvement in limb support & stability in the first 3 trials with no further improvement.

Pai et al. [57]

n = 67, 44♀, 72 ± 5.5y. Community dwelling, healthy.

Slip (low friction moveable platform; slid up to 90 cm forward at foot contact).

24 slips in 37 gait trials.

Retrospective falls data 12mo preceding & 12mo prospective following the session.

Reduction in falls 12mo post-session compared to 12mo pre-session (15% compared to 34% incidence).

Pai et al. [63]

3 groups tested +6, +9 & +12mo respectively: +6mo: n = 24, 13♀ 74.6 ± 5.8y; +9mo: n = 23, 15♀ 71.8 ± 5.5y; +12mo: n = 26, 19♀ 72.0 ± 4.7y. All community dwelling, healthy.

Slip (low friction moveable platform; slid up to 90 cm forward at foot contact).

24 slips in 37 gait trials.

Proactive & reactive stability (measured at touchdown of the to-be-perturbed step & the first recovery step respectively). % of falls following the lab perturbations.

Falls reduction from 42.5% to 0%. 0%, 8.7% & 11.5% of participants at the +6mo, +9mo & +12mo slips respectively fell. Stability improved & was better at all time points than the first slip.

Parijat and Lockhart [58]

Training: n = 12, 71.24 ± 6.82y. Control: n = 12, 74.18 ± 5.82y 12♂ 12♀. All community dwelling, healthy.

Pre-post: slippery surface. Training: slip (moveable platform; 30 cm at 1.2 m/s forward) with ±20% velocity based on ability.

12 slips in 24 gait trials.

Slip distance & peak sliding heel velocity pre- & post-training on the slippery surface. % of falls following the lab perturbations.

Falls reduction on slippery surface from 42% (pre) to 0% (post). The reduction in slip distance & peak sliding heel velocity was greater in the training group.

Sakai et al. [66]

n = 45, 26♀, 71.4 ± 3.6y. Community dwelling, healthy.

Treadmill decelerations at heel strike during walking at 2 km/h. 50% reduction in belt speed lasting 500 ms.

20 sudden treadmill belt decelerations at heel strike during 5mins of walking.

Peak forward & backward sacrum accelerations (accelerometer) within 1 gait cycle post-perturbation (average of first & last 10 perturbations).

Mean peak sacrum accelerations were lower in the final 10, compared to the first 10 perturbations.