Impact of system maturity/design and test settings on the evaluation of assistive tongue computer interfaces

When assistive computer interfaces are introduced or evaluated, they are often compared to existing systems. Such comparisons are challenging since many parameters affects the recorded typing efficiency. For tongue computer interfaces such parameters include: Degree of oral integration of the system, whether an activation unit is fixed to the tongue with glue or through a piercing, the design of the activation unit, whether testing relates to the time to execute a command or to type using the whole alphabet and whether errors are allowed. This study describes how different design and test features of tongue computer interfaces affect the typing efficiency of tongue computer interfaces. An experiment was performed to study the effect of activation unit design, and examples from literature on how the recorded efficiency of three different tongue computer interfaces relate to maturity and test setup is combined into an tongue computer evaluation-matrix. The results show great dependency on system maturity and test set-up, where only approximately 10% of the reported typing efficiency in the immature systems is preserved in the mature systems. This suggest that performing, interpreting and especially comparing system test may not be unambiguous. The results suggest the relevance of similar studies on a maturity-efficiency-matrix for other assistive computer interfaces such as e.g.brain computer interfaces.