BCI-based rehabilitation games are appealing for cerebral palsy because engagement drives training effort, and training effort is what produces benefit. What is not clear is how the systems should be designed, and the authors identify two specific gaps.

The first is how people with CP actually experience a sense of control through a BCI. Agency is the whole point of the interaction, and it is subjective: an interface can be measurably accurate and still not feel like the person is the one doing it.

The second is more delicate. These games assist the player, and how that assistance is experienced determines whether the exercise still counts as one's own effort. Too little help and the task is discouraging; too much and success stops belonging to the player. That is a design question no accuracy metric answers, which is why asking the users is the method here.