*2.2. Experimental Design*

Subjects performed a repetitive screwing task with their dominant hand on a wooden plate that was positioned in front of them. The height of this plate was individually adjusted, so that the middle of the vertical screwing rows was aligned with the elbow height of the subjects while they were standing upright, with the hand palms facing anterior and the thumbs pointing lateral. For 60 min, subjects screwed 72 screws (4.5 × 45 mm, T-Start T20, 4 g, SPAX International GmbH & Co. KG, Ennepetal, Germany) into the wooden plate (multiplex, 30 mm thick), which was split into 12 vertical rows of six screws each. With their non-dominant hand the subjects positioned the 6 screws, taken from the container that was positioned close to their non-dominant hand, and screwed them in one after another with their dominant hand using a T-handle screwdriver (e.g., T-handle 336, T15, handle cross size 80 mm, shaft length 200 mm, 162 g incl. 3 g bit, Wiha, Germany; Figure 1), after which they pressed a buzzer (phase 1). Then, the subjects changed tools and fastened the same 6 screws with their dominant hand using a torque screwdriver (7443 pistol, 232 g incl. 3 g bit, Wera, Germany; Figure 1), after which they pressed a buzzer again (phase 2). These actions were repeated 12 times until all 12 vertical rows were filled with 6 screws. Due to using torx head screws, the amount of forward directed force was low. The torque screwdriver used for fastening released a click at a predetermined torque of 5 Nm. The buzzer was sampled along with the muscle activity recordings.

Subjects visited the laboratory on three days, separated by 2 to 7 days. At day 1, subjects performed for a period of ~10 min, during which they familiarized with, i.e., got acquainted with, the task. Subjects were considered familiar with the task when they were able to perform screwing and fastening with the two tools provided and work at the predetermined work pace. Following familiarization and at days 2 and 3, the subjects were prepared for the measurements, performed reference measurements for electromyographic (EMG) normalization, and performed the one-hour task.

The work pace of the subjects was controlled using visual feedback, i.e., a vertical bar representing a timeline showed the time that was left to fulfill screwing (phase 1) and fastening (phase 2) of each of the 12 rows. Work pace was predetermined based on the standardized, predetermined motion-time measurement system MTM [18], a system often used in industrial production planning. For both phases together we chose pace MTM-85 to allow all subjects to fulfill the task in the given time without developing muscular fatigue. This pace corresponded to 242 s for screwing (6 screws; phase 1) and 28 s for fastening (6 screws; phase 2), adding up to 270 s for one vertical row. Since we controlled the amount of work and the work pace and since we observed that all subjects screwed in the screws completely, we can conclude that performance was equal within subjects across days and also across subjects.

### *2.3. Experimental Design*

#### 2.3.1. Muscular Activity

Electrical activities of the following four dominant superficial muscles of the hand-arm system that are involved in grasping and screwing, as verified in pilot measurements and as deduced from previous studies [19,20], were measured: The M. triceps brachii, M. biceps brachii, M. flexor carpi radialis, and M. extensor digitorum. Their electrical activity was recorded using surface EMG. We placed pre-gelled Ag/AgCl surface electrodes (35 × 26 mm, 15 mm active area diameter, KendallTM H93SG ECG Electrodes, Covidien, Zaltbommel, the Netherlands) in a bipolar configuration with an inter-electrode distance of 26 mm (center-to-center) on the muscle belly, according to Criswell [21]. A ground electrode, used to equalize the electrical ground level of the measurement device to the subject's electrical level and to minimize electromagnetic interferences, was placed on the seventh cervical vertebra. Prior to electrode placement we shaved the skin, in case of body hair, and prepared it with an abrasive paste (Skin Prep Gel, Nuprep®, Aurora, USA).

EMG signals were differential amplified, analogue filtered (2nd order high pass filter, –3 dB at 4 Hz; 11th order low pass filter, –3 dB at 1300 Hz), sampled (4096 Hz), analyzed, and stored using a combined data analyzer and logger (PS11-UD, THUMEDI® GmbH & Co. KG, Thum, Germany; overall CMRR > 96 dB; overall effective noise < 0.8 μV RMS; linearity type ±0.25 dB at 20–1100 Hz). The device real-time transformed the data into the frequency domain (1,024-point Fast Fourier Transformation, Bartlett-window, 50% overlap) and digitally high-pass filtered the signal (11th order; –3 dB at 20 Hz). Power line interferences (50 Hz and its first seven harmonics) were removed by replacing it by the spectral values of a 4-Hz wide bandwidth around its center frequency by means of both spectral neighbors. The root-mean-square (RMS) of the electrical activity (μV) and median power frequency (MPF (Hz)) were real-time calculated from the power spectrum (250-millissecond moving window with 50 % overlap) and stored synchronously to the raw data by the PS11 device. An example recording of the RMS of the triceps brachii is provided in Figure S1 (Supplementary Material).

#### Normalization

Prior to the experimental task, we collected EMG during four 15-s submaximal reference voluntary contractions (RVC) with fixed force levels (Table 2). Each RVC targeted one of the four muscles, biceps brachii, triceps brachii, extensor digitorum, or flexor carpi radialis. The RVCs were measured using a self-developed measuring device in which the subject took a standardized position. With the upper body in an upright position, the forearm was placed horizontally and the upper arm perpendicular to the forearm. For determining the force of the biceps brachii and triceps brachii muscles, the force cell was positioned underneath the cushion below the distal end of the forearm. For determining the force of the extensor digitorum and flexor carpi radialis in the forearm, the force cell was positioned underneath the metacarpal bones in the hand. A monitor was connected to the force cell to give subjects visual feedback on the force level of the muscle contraction. Every RVC was followed by a recovery period of ~1 min. Electrical activity of the muscles was recorded during the RVCs and the middle 10 s of a steady-state period was used for EMG normalization. All RMS values were expressed as a percent of the electrical activity during the RVC, i.e., reference voluntary electrical activity (%RVE; Table 2), by dividing the experimental electrical activity level (μV) by the reference electrical activity level (μV) and multiplying by 100%.


**Table 2.** Descriptions of submaximal reference voluntary contractions (RVC) of the four target muscles, with corresponding force levels at which subjects had to sustain the RVC.

<sup>1</sup> The force levels [N] were determined during pilot measurements among five subjects to correspond moderate force levels of ~40% maximal force.
