This is the second update from a project supported by the OpenBCI Discovery Program. Click here for details on how to apply.
Much has happened on the Koalacademy project over the past six months, since our first OpenBCI Community post. If you haven’t already given the original article a read, here it is. You can also read about the project directly on the NeurAlbertaTech (NAT) website here (NAT is the neurotech innovator network that spawned the Koalacademy project team).
Koalacademy is a BCI-optimized language learning platform that will use live brain activity during studying to predict whether learning will be successful or not, relaying that information back to the user to improve learning rates.
The following are some big picture milestones that the Koalacademy project has passed, as of late:
A. Continued Development
With the bones of the architecture worked out we are continuing development on multiple fronts:
- Implementing an embedded Unity game served by our React website, which features a 3D rendering of a Koala interacting with word associations.
- Creating custom client-side preprocessing pipelines for EEG.
- Flexibility to connect to other OpenBCI hardware (besides the Cyton-Daisy), such as the Ganglion (4 channel) and the Cyton (8 channel).
- Integrating social media and community features.
B. Academic Research Study
With the core functionality down we are now partnering with some wonderful researchers to provide rigid validation of the Koalacademy platform.
We will be utilizing a number of Cyton-Daisy headsets (a small, a medium, and a large), all with the integrated ThinkPulse™ Active Electrodes in order to collect data.
A methodological study showing that the web framework we have developed is capable of presenting stimuli and streaming brain data in a synchronized manner. Essentially, we will be demonstrating that Koalacademy is comparable to other, more traditional and previously validated means of acquisitioning data from the OpenBCI. We will be comparing a control condition, wherein the Cyton-Daisy data is streaming to an external Raspberry Pi 4 (RPi4) + triggered by a light dependent resistor attached from the RPi4 to the monitor to detect the stimulus (presented through the Koalacademy framework). We will then compare the ERPs derived from both conditions with participants completing a traditional oddball task.
A within-subjects design demonstrating that the machine learning model we generate from a user’s brain and behavioural data (synchronized to the onset of word association pairs) is internally valid. Wherein, we are validating whether a model is able to successfully predict subsequent recall in real time. More details on this study will be provided in our next OpenBCI Community post.
C. Faculty Support
We have had the opportunity to connect with several very interested AND supportive faculty at the University of Alberta:
Associate Professor, Faculty of Rehabilitation Medicine, Department of Communication Sciences & Disorders
Canada Research Chair in Neuroscience of Literacy
Generously providing her laboratory space, funding, and expertise throughout both Study I and Study II.
Professor, Faculty of Arts, Department of East Asian Studies
Directly providing invaluable support with respect to the Mandarin and the psycholinguistic aspects of Study II.
Associate Professor, Faculty of Kinesiology, Sport, and Recreation – Academic Programs
Provided grant writing support and endorsement.
As team lead, Eden Redman has personally received an undergraduate grant from the Branch Out Neurological Foundation (BONF) that will serve to support both validation studies happening this Spring and Summer (2021). BONF is an Alberta granting agency that funds innovative studies outside of the realm of typical research, with a focus on disseminating research findings in an impactful manner.
E. Opportunities with NeurAlbertaTech
NeurAlbertaTech, the parent organization of Koalacademy, is hosting an accessible BCI hackathon THIS Summer, with over $10,000 in prizes! Find out more about the hackathon and how to apply by visiting our website (https://neuralberta.tech/event/nathack).