Media Player for the severely impaired using brain-computer interface (BCI) and open data

[mozfest-bot]

[ ID ] 492faa56-99f1-417a-8df7-885ec92661e9

[ Submitter's Name ] Abhiram Ravikumar [ Submitter's Affiliated Organisation ] Mozilla India [ Submitter's Twitter ] abhi12ravi

[ Space ] science [ Secondary Space ] exhibit

[ Format ] learning-lab

Description

Our session will give the audience an overview of brain-computer interface. The participants will have access to the media player built by us which caters to the severely impaired people. To begin with, participants will be introduced to the lobes of the human brain and their functions, followed by the EEG bands and their characteristics. They will then learn about the various types BCI designs- invasive, partially invasive and non-invasive - along with the neuroimaging techniques such as ECoG, EEG, etc. associated with each. The input signals- which mirror imagined hand movements- are then pre-processed, significant features are extracted and classified into right and left hand movement. This is done using machine learning algorithms like Common Spatial Filter and Linear Discriminant Analysis. At the end, the participants can use the application (https://github.com/bhoom10/bci-media-player) and give us feedback. The larger goal is to build a community of developers for the project.

Agenda

While brain computer interface is a developing field, it's still in its infancy. We hope to educate participants about the field and the obstacles faced by those with severe disability who require BCI to carry on with their lives. It'll be done through a series of activities and games that will create awareness. Following this, we wish to explain our research in the field. We used openly available BCI datasets (http://www.bbci.de/competition/iii/#datasets) to train our classifier. The research is based on using motor imagery of the left hand and right hand to operate a media player using a binary-like combination. This dataset is pre-processed using a bandpass filter and then CSP. LDA is used for classification of left and right hand signals. Results achieved are compared with the verified results for accuracy. This input is then verified via a regular expression and used to control the media player.

Participants

We propose to bring a sense of awareness among participants about the difficult lives led by people with severe disabilities. We wish to expose the participants to the failure of current assistive technologies in order to carry on with their normal lives. This will be done through a set of simple yet effective exercises and games. This includes tasks such as walking across a room with a blindfold, communication through sign language, sitting still for a while and the like. Group activities such as coordinating a small dance with a blindfold and guiding a blindfolded participant through a series of obstacles will also be conducted.These activities can be designed to accommodate teams of varying sizes based on the size of the audience. These activities allow participants to place themselves in the shoes of the mentally challenged and understand their thought process.

Outcome

The main outcome of our activities is to create awareness about the challenges faced by people with disabilities, and also explore the solutions that are currently in development. Our tool, built using openly available data - is available on GitHub (https://github.com/bhoom10/bci-media-player) for testing, although several modules are yet to be built. We hope that the increased awareness will encourage participants to come forward and join our efforts to empower people with severe disability. Participants of the session will be able to interact with the application while developers shall understand how the app is built so that they can continue the project.
We wish to make the development of the media player a crowd-sourced effort. At the same time, we also wish to encourage people with disabilities to adopt the MP3 player and make the most of it to empower themselves.

[thornet]

cc @cubicgarden for neuro-diversity interests

[abhi12ravi]

Hi @arlissc, thanks for the acceptance. However, we haven't received a confirmation e-mail yet. Please let us know what to do next, thanks!

Cheers, @bhoom10 & @abhi12ravi

[arlissc]

Hi @abhi12ravi - the offer of a stipend to MozFest was the indication that your session was accepted. Our apologies if this was not clear.

[abhi12ravi]

Hi @arlissc thank you for the acceptance and the stipend. Unfortunately, we won't be able to make it to MozFest this year. Thanks for all your motivation and support. I hope we can turn this project into reality with the help of Mozilla Science Lab. Looking forward..

Cheers, @bhoom10 & @abhi12ravi