Motor Movemet Passkey Project
Biometric security techniques such as fingerprint or facial recognition are gaining popularity due to the high degree of security they provide. Our BCI team has built an alternate biometric system that uses brain data recorded from an EEG to authenticate users. The system uses motor movement passkeys to learn the brain signals of a given user such as opening a hand and then moving both feet which are chained together to unlock a device. These movements produce brain signals in the motor cortex that are picked up by the EEG and analyzed by a binary neural network to grant access to the device. We believe this research can be applied in fields such as banking or government where security is of extremely high importance. The experiment was conducted with data from Physionet and the code as well as the datasets can be found below.
Brain-Computer Interface Team
A brain-computer interface (BCI), sometimes called a neural control interface, is a direct communication pathway between a brain and an external device. Our brain-computer interface team designs non-invasive BCIs for live data acquisition with the use of wireless electroencephalographs (EEG). Currently, the team is working on optimizing the 16 channel and 32 channel electroencephalograph systems for application-specific use. Implementations of the BCIs in cryptography, neurorobotics, and mental illness are under development. The team consists of students from the faculties of engineering and computing.
Cognitive Engineering Research Team
Cognitive engineering is the application of cognitive psychology to the design and operation of human-machine systems. In other words, it's the study of how we can build systems to better fit the needs of users. Our research team works to synthesize existing data with the products of our brain-computer interface team to draw new conclusions in the study of neural systems. As a student research team, we have a unique perspective and this enriches the work that we do. We feel that studying these issues will provide insight into the fundamental problems facing brain-machine interfaces and help to better integrate BCIs into real applications. The team consists of students from the faculties of science and engineering.
Queen's Cerebral Language Innovation has presented research work at various conferences including the NeuGeneration neuroscience conference and the Computing Student Association conference. This year in NeuGeneration 2021, QCLI ran an EEG encryption workshop. In the workshop, delegates had the opportunity to encrypt and decrypt their own message with a specific set of pre-recorded motor movements from the brain. Using the biometric encryption algorithm we created, the delegates were able to see the process of secure brain signal transmission from start to finish. Our presentation at COMPSA 2020 was also a great way to stay connected to the computing community and recruit great candidates for our teams.
The most powerful artificially intelligent language model ever created was recently developed by OpenAI, an open-source AI company founded by Elon Musk and Sam Altman. QCLI was recently approved by OpenAI's selection process and is now granted access to the API for the GPT-3 neural network. We are currently developing implementations of this powerful network for EEG data analysis and language reconstruction. There are several other networks provided by OpenAI for mechanical control and image generation. These networks and models will be integrated into some of our other brain-computer interface applications mainly related to neuromechanical control.