Physical workload during use of speech recognition and traditional computer input devices

Musculoskeletal symptoms among computer users are frequently found. The aim was to investigate the musculoskeletal workload during computer work using speech recognition and traditional computer input devices (keyboard/mouse). Ten experienced computer users (nine female, one male) participated. They performed three different computer tasks: (1) text entry and (2) text editing of a standard text and (3) a self-selected work task. These tasks were performed twice using speech recognition and traditional computer input devices (keyboard/mouse). Additionally, a task consisting of reading aloud of the standard text was performed. Surface EMG from the forearm (m. extensor carpi ulnaris, m. extensor carpi radialis), the shoulder (m. trapezius) and the neck extensor muscles was recorded, in addition to the voice-related muscles (m. scalenii, m. cricothyroideus). Using speech recognition during text entry and text editing reduced the static muscle activity of the forearm, neck and to some extent the shoulder muscles. Furthermore, tendencies to longer periods of muscle activity pause (relative time with EMG gaps) in the forearm and shoulder muscles were found. This was seen at the expense of a tendency to an increased static activity and a decreased relative time with EMG gaps in m. cricothyroideus. Finally, during use of speech recognition the hand was tied to the keyboard/mouse for a shorter period of time, while the eyes were viewing the screen for a longer period of time compared to the condition with traditional computer input devices. It is recommended to use speech recognition as a supplementary tool to traditional computer input devices.