The EarOpener is a project I was part of while interning with the Cognitive Artificial Intelligence Group at the Donders Institute for Brain, Cognition and Behaviour, Nijmegen, the Netherlands.
I joined this project because I was fascinated about the possibilities that Brain Computer Interfaces opened up and Neurofeedback sounded like a very intriguing procedure. I was coming from another research project on rodents and I wanted to switch to human subjects and EEG. It was also the perfect opportunity for hands on experience with the cognitive science of language.
Second language (L2) perception is influenced by the perceptual organization of native language (L1). Even after intensive behavioral perceptual discrimination training, L2 speakers might never achieve native level proficiency. This has profound implications ranging from integration of migrants to job success in a new country.
The EarOpener was a project that explored the possibility of using neurofeedback based on two event-related potential (ERP) components: the mismatch negativity (MMN) and P300, with the goal of helping L2 learners discriminate between sounds for which they lost sensitivity during childhood.
My job was to come up with a project proposal that will advance the EarOpener work in an unexplored field. After intensive study of the literature and exploring options, I decided on the following approach: using neurofeedback to shift a categorical boundary for Dutch learners of English. Under the supervision of Dr Makiko Sadakata, I prototyped, designed and conduced the experiment, I gathered, analysed and presented my results in a research paper. I designed a game in order to elicit intrinsic motivations from our participants. Classification of brain responses was done by means of quadratically regularized linear logistic regression classifiers (Bishop 2009). We did not find conclusive evidence that our approach could generate a shift a phonetic categorical boundary.