Fig 1.
Control methods and devices trialled.
Conventional control mechanisms were trialled using standard tablet and smartphone (A, B). Subjects were required only to move a cursor along a single vertical path, full-range, and then to an indicated vertical level (they were not tested on playing the underlying game). B shows software used for assessing swipe, with varying cursor size. There was no improvement in accessibility using a larger cursor. The novel control mechanism (C) is a wireless grip-force sensor that detects both finger-flexion and extension movements, the latter assisted by a fingerstrap holding the device within a partially-extended hand. Control software for C entailed moving a circle in a vertical plane towards a target star. Cursor and target stimuli dimensions and contrast are similar between all methods.
Fig 2.
Subjects were asked to move the cursor three times up and down the longest vertical path, as well as to a position level with an indicated adjoining horizontal path.
Fig 3.
This shows numbers of arm-paretic stroke patients screened, excluded and recruited, and reasons for exclusion.
Table 1.
Patient demographics and baseline clinical characteristics.
Fig 4.
Control ability using conventional versus novel controllers.
A: Proportions of patients achieving each level of cursor control (0–3) for each of the four conventional, and one novel (grip), control mechanism. Results are stratified according to severity of arm weakness (using Short-Fugl-Meyer score of the arm). B: Performance error on 2-minute tracking task controlled by grip-control, plotted against arm disability. A small trend towards less error with greater ability is non-significant whether or not the one outlier is included (dashed-line) or not (continuous-line) (p>0.1 for both)–indicating that tracking accuracy is largely independent of standard arm-function scores.