Exploring task design for cognitive commands and prerequisites of accessibility with a visual Imagery task for spontaneous EEG
Master thesis
Published version
Permanent lenke
https://hdl.handle.net/10642/8671Utgivelsesdato
2018Metadata
Vis full innførselSamlinger
Sammendrag
I have explored skills required for end-users to operate a BCI communication system, attempted to define
optimal cognitive commands for task design with Stimulus-Response compatibility, and discussed how
hardware and system design may support accessibility. I have performed experiments with an Emotiv EEG
headset and a game interface to find and the participants have reported perceived Workload with a NASA
Tlx form. I have assessed the effectiveness and user experience of spontaneous EEG, by testing training
data with different cognitive tasks and compared them with number of classifications per session, and the
score achieved. There are benefits in user motivation and workload if we can adjust tasks to skill level and
customize stepwise skill acquisition to the individual user. I have hypothesised that a trade-off in dataset
similarity may outweigh the benefits of a more accessible workflow. There is no significant evidence of a
difference in tested efficiency or accuracy between using visual imagery and visual perception as training
data with visual imagery gameplay task. There is no significant evidence of a difference in perceived
workload between playing with different sets of training data. I discuss how further development in machine
learning models may fascilitate even further to a desired accessibility with cognitive commands
Beskrivelse
Master i universell utforming av IKT