Neuromuscular control of gait stability in older adults is adapted to environmental demands but not improved after standing balance training

Balance training aims to improve balance and transfer acquired skills to real-life tasks and conditions. How older adults adapt gait control to different conditions, and whether these adaptations are altered by balance training remains unclear. We investigated adaptations in neuromuscular control of gait in twenty-two older adults (72.6 ± 4.2 years) between normal (NW) and narrow-base walking (NBW), and the effects of a standing balance training program shown to enhance unipedal balance control in the same participants. At baseline, after one session and after 3-weeks of training, kinematics and EMG of NW and NBW on a treadmill were measured. Gait parameters and temporal activation profiles of five synergies extracted from 11 muscles were compared between time-points and gait conditions. No effects of balance training or interactions between training and walking condition on gait parameters or synergies were found. Trunk center of mass (CoM) displacement and velocity (vCoM), and the local divergence exponent (LDE), were lower in NBW compared to NW. For synergies associated with stance of the non-dominant leg and weight acceptance of the dominant leg, full width at half maximum (FWHM) of the activation profiles was smaller in NBW compared to NW. For the synergy associated with non-dominant heel strike, FWHM was greater in NBW compared to NW. The Center of Activation (CoA) of the activation profile associated with dominant leg stance occurred earlier in NBW compared to NW. CoAs of activation profile associated with non-dominant stance and non-dominant and dominant heel strikes were delayed in NBW compared to NW. The adaptations of synergies to NBW can be interpreted as related to a more cautious weight transfer to the new stance leg and enhanced control over CoM movement in the stance phase. However, control of mediolateral gait stability and these adaptations were not affected by balance training.

79 characterize effects on the neuromuscular control of gait and of adaptations to narrow-base 80 walking. 81 Methods

82
The methods described here in part overlap with our previous paper [24], as data were 83 obtained in the same cohort.
84 Participants 85 Twenty-two older (72.6 ± 4.2 years old; mean ± SD, 11 females) healthy volunteers 86 participated in this study. Participants were recruited through a radio announcement, contacting 87 older adults who previously participated in our research, flyers and information meetings.  176 Therefore, in the current study we fixed muscle weightings between conditions and time- Five muscle synergies were extracted with a fixed muscle weighting matrix H (Fig 6) and 226 activation profiles per individual per condition and time-point (Fig 7). This accounted for 227 87±2% of the variance in the EMG data.  (Fig 7). This may reflect a slower weight 296 acceptance by the new support leg, possibly related to the lower activation peak during push-297 off observable in synergy 1.

298
The CoA of the activation profiles was different between conditions but was not affected 299 by training. Narrowing step width led to an earlier CoA of the activation profile associated with 300 dominant leg stance (synergy 1) and delayed CoAs of the activation profile associated with 301 dominant and non-dominant leg heel strikes (synergies 4 and 5). Earlier CoA in the dominant 302 leg stance phase appears to be a consequence of the reduction in activation during the second 303 peak of the activation profile (Fig 7). This reduction in activation would reflect a decrease in 304 muscle activity related to push-off and possibly reflects a more cautious gait. The earlier CoA 305 of the activation profile associated with heel strike reflects a more sustained activation 306 following a slower build-up (Fig 7). Again, this may be related to a more cautious walking but 307 also to active control over CoM movement during the stance phase. The latter is supported by 308 the fact that muscles that would contribute to mediolateral control, specifically tibialis anterior, 309 peroneus longus and gluteus medius are part of these synergies. To check that changes in CoA 310 and FWHM of the activation profiles were not due to changes in duration of gait phases, we 311 assessed single support and double support times as percentages of the stride times and no 312 effects of condition were found.

313
We studied effects of a balance training program of only 3-weeks. For transfer of acquired 314 skills to a new task, it may be necessary that a high skill level is achieved and possibly more 315 than 3 weeks are needed. Improved gait parameters were reported after 12 weeks of balance 316 training [5]. Therefore, a longer duration of training might have led to changes in mediolateral 317 gait stability.

318
In conclusion, older adults adapted mediolateral CoM kinematics during gait to narrow-319 base walking and this was associated with changes in synergies governing the activation of leg 320 muscles. However, we found no evidence of a change in control of mediolateral gait stability,