*2.2. Procedure*

The baseline assessment comprised an interview that focused on (1) overall medical health, (2) anthropometric and socio-demographic characteristics (age, sex, height, weight, BMI, and Spanish foot size), (3) chronic disorders and illnesses (e.g., arthritis, diabetes mellitus, and musculoskeletal disturbances), and (4) recreational activities. A senior podiatrist also assessed each rider with a complete physical clinical evaluation and measured anthropometric variables such as height, weight, and body mass index.

Before testing commenced, the participants were given a two-minute acclimation period to drive on the motorcycle simulator. During this period, participants drove on the motorcycle simulator to adjust to the surroundings and ensure the shoes were comfortable. Each motorider wore their own shoes with a flat thin rubber sole.

In the simulator CKU28® (Mecanitzats Muntada SL, Manlleu, Spain), the riders drove the Circuit of the Americas race (Austin, TX, USA) in which professional sport motorcyclists compete [17]. During the simulation, riders drove the race course, making the same movements as real circuit motorcycle riders would need to make. Riders had to accelerate to maximum speed and then, when taking a curve, decelerate. They needed to raise and lower their bodies on the motorcycle simulator to the degrees needed to complete the course. The simulator had a maximum inclination of 60◦ and riders rested their knees on the ground just as they do on real motorcycles. Riders completed the race at approximately 250 km/h in straightaways and at decreased speeds in curves over a period of 20 min of Racing for each CFI test.

We evaluated four types of CFI: (1) only polypropylene (58◦ Shore D), (2) selective aluminum (60 HB Brinell hardness) in the metatarsal and first hallux areas and polypropylene elsewhere (58◦ Shore D), (3) ethylene vinyl acetate (EVA) (52◦ Shore A), and (4) standard EVA (25◦ Shore A) as control. We tested each rider with each type of CFI and the order of CFIs was randomly assigned using a random number table. The order of CFIs was the same across all riders.

Before testing, each motorcyclist was situated in a calm place to fix surface electrodes (positive, negative, and earth interfaces) with a built-in preamplifier on the skin over six muscle areas on the right leg: (1) fibularis brevis, (2) fibularis longus, (3) tibialis anterior, (4) gastrocnemius lateralis, (5) gastrocnemius medialis, and (6) soleus. The belly of each muscle was marked on the skin with a permanent marker, swabbed with a 70% alcohol solution and then shaved. The skin was then abraded with skin sandpaper to reduce the electrical impedance to less than 5 kX. All active electrodes were placed 2 cm apart, parallel with the alignment of the muscle fibers, and secured with hypoallergenic tape. A stretch "underwrap" bandage was placed over the electrodes to prevent electrode movement [18].

The precise locations of the EMG electrodes were as follows:


For normalization purposes, before performing each trial, the participant was instructed on how to perform a maximum voluntary isometric contraction (MVIC) for each muscle based on a technique described by Yang and Winter [19]. Three MVICs were recorded for each muscle within a recording time of eight seconds. This allowed time for the participant to build up the contraction and for the tester to stabilize the foot. Thus, the MVIC was measured over a four-second window only. The participants were instructed to exert a maximum effort against the resistance of the tester and were given verbal encouragement while doing so. The rationale for this test is to normalize the maximal amplitude of a submaximal task, such as walking or riding a motorcycle, to the amplitude of the MVIC produced by the participant. One of the known limitations related to the reliability of the MVIC is the participant's ability to produce an MVIC sincerely [19]. Although the participant's effort can vary between each contraction, using the MVIC is better than not using a normalization technique at all [20]. To reduce

the effect of muscle fatigue between succeeding contractions, participants were given one-minute rest periods [21].

The EMG electrodes were inserted into a lightweight backpack EMG interface unit (model MWX8). The interface unit was wired to the filtering hardware by an overhead pulley system. This was connected to a personal computer where EMG information was recorded using DataLINK (v.5.0) (Biometrics Ltd., Ladysmith, VA, USA).

We measured the signal activity patterns with surface electromyography (EMG) sensors (SX230) (Biometrics Ltd., Ladysmith, VA, USA), using the protocol design of SENIAM to comply with a non-invasive evaluation of muscles [12]. Electrode specifications in this study were as follows: amplification, single differential; inter-electrode distance, 2 cm; contact sensors, two 0.37 × 0.20 × 0.06 cm silver bars; preamplifier gain, 1000; input impedance, >10,000,000 MΩ; and common mode rejection ratio, >96 dB. The main amplifier unit feature was a gain of 1000-fold and frequency response of 20–460 Hz. Surface electromyography data were sampled at 1000 Hz and recorded by a computerized data logger (MWX8 Biometrics Ltd., Ladysmith, VA, USA).

We chose the 1 kHz data sampling rate because raw EMG signals were collected, converted with DataLINK, and recorded prior to producing the filtered/rectified data. This ensured that the peaks and onsets of activity were reproduced with subsequent data, and enabled the verification that movement artifacts were minimal [18]. Electromyography signals recorded during the selected times were filtered with a band pass filter (bandwidth of 20–460 Hz) [21,22]. EMG parameters such as frequency content and waveform shape may require sampling rates more than twice the Nyquist rate [23] for accurate interpretation. However, for the analysis of amplitude and timing parameters, over-sampling is unnecessary [24].

The average rectified value (ARV) was calculated. The total ARV and ARV during riding for each muscle and each participant were calculated and normalized by the peak ARV for each participant. To identify the muscular activation phase, the onset and offset of surface electromyography were assessed. Phases of muscular activation were defined by the period when the signal was above a threshold of 25% of the peak ARV for each muscle and participant [25,26]. The signal activity patterns related to the LLM were registered for each CFI test during the last five minutes of the simulated race.

For each riding, "raw" recordings (without bandpass or RMS/envelope filtering) were taken to compare with filtered recordings. The signals were then filtered during testing for each participant.

We calculated the root mean square (RMS) of each signal for each subject, and used a filter bank of 10 wavelets. We also computed center frequencies that characterize the wavelets [27]. These operations yielded a time series of EMG intensities for each wavelet (wavelet domain). These time series were re-sampled at 200 Hz and divided by the square root of the sum of the intensities of the data points in all wavelet domains. This yielded intensity patterns, which we used in blocks for further analysis.

During the collection and analysis of EMG tracings, the operator was blinded to the condition being evaluated. Muscle onset was determined by visual observation as the time when a rapid and continuous increase in voltage occurred that was unrelated to movement artifacts or noise interference.

Lastly, all EMG electrodes were placed on the muscles of the right lower extremity of each motorider, so the right curve means that the right lower extremity is touching the floor with the knee and applying force against the footpeg, while the left curve means that the same lower limb must move towards the motorcycle body part to maintain stability.
