Movement Cycle

The movement cycle was defined as the interval between two consecutive maximum knee flexions of the right knee. The maximum knee flexion is an event occurring both in bicycling and in walking. Therefore, this approach facilitates the comparison between the two types of movement. However, this definition deviates from previous walking studies which used the heel strike as the boundary of the gait cycle (Gwin et al., 2011; Wagner et al., 2012; Seeber et al., 2014). The maximum knee flexion occurs between toe off and heel strike. Power modulations locked to the movement cycle based on the knee flexion precede the ones based on the heel strike by approximately 30% of the movement cycle.

To avoid confounds due to acceleration or deceleration, we solely incorporated cycles deriving from the continuous movement condition and from the period between 2 s after movement start and 2 s before movement stop. This resulted in an average of 202.1 ± 22.2 movement cycles (range 160–240). Time-frequency spectra were calculated per movement cycle with a sliding Hanning window. The window had a length of 0.5 s and was moved in steps of 0.01 s. Individual spectra were interpolated to the length of the longest cycle and subsequently averaged. This procedure standardized the same time frame for all subjects and conditions. Power within the movement cycle was normalized by the temporal average of the entire cycle for each frequency.

Rhythmicity of power modulations within the movement cycle was quantified by the gait phase modulation measure introduced by Seeber et al. (2014) and modified by Trenado (2015). This index can take values between 0 and 1 reflecting the similarity of the power modulation within the movement cycle with a sinusoid. The calculation is based on two periods of the sinusoid representing the modulation with every single leg movement within the movement cycle. The frequency and peak value of maximal sinusoidal modulation within the 24–40 Hz range was computed for every subject and condition. The choice of the frequency range of interest was based on prior studies (Wagner et al., 2012; Seeber et al., 2014). Peak values of the modulation measure did not differ between conditions, i.e., power modulation within the movement cycle had the same similarity with a sinusoid in bicycling and walking (Bicycling: M = 0.78, SD = 0.13, Walking: M = 0.83, SD = 0.10, paired t-test, p > 0.05). The resulting frequency was further used for fitting a sine wave to the corresponding power modulation. Again, a two-period pattern of the sine wave was assumed. This allowed comparing both phase and amplitude of rhythmic power modulations between bicycling and walking.