*4.2. MuscleSound® Score (2015)*

Two of the researchers involved with this study were trained to capture ultrasound images on the vastus lateralis using a portable ultrasound machine. They practised this technique to achieve acceptable reliability with repeat images and were assigned to the study roster so that each participant was scanned by a single technician over the duration of their study involvement. Thereafter, on each occasion that a muscle biopsy was performed, ultrasound images were captured using the same machine (Terason T3000, TeraTech Corporation, Burlington, MA, USA) and by the same technician. Five ultrasound images were captured at each time-point on the vastus lateralis of the contra-lateral leg, tracking the location of the specific incision on the biopsied leg (Figure 1: U1–U4). A further five images were captured on the contra-lateral leg at the site of the original (baseline) assessment (Figure 1: U1). The protocol followed MuscleSound® guidelines with images captured on the transverse plane at a depth of 4 cm and a gain of 45 with the muscle relaxed. The

transducer head was manipulated to achieve a bright fascia, which defined the muscle boundary for the region of interest. Images were then uploaded to the MuscleSound® software (v.2015, MuscleSound®, LLC, Denver, CO, USA) and processed according to their proprietary protocols. The MuscleSound® score (0–100) was subsequently provided in arbitrary units (a.u.), noting that such scores were provided in bands of 5 a.u. [5]. A single score used in the statistical analysis was obtained by averaging the score from each of the five images at each site.

## *4.3. Estimated Fuel Level (2017)*

The ultrasound images were re-analysed by MuscleSound® using an updated protocol titled "estimated fuel level". Details of this proprietary process were not published, but are described as "placing an image in context of the maximum (100) and minimum (0) points of glycogen obtained from a bank of images captured for a specific participant [13]". We have described these values as "points" to distinguish them from the original metrics (described as a.u.).
