Note: the format did not transfer well, I'd be happy to share the pdf if needed!
Assistive Environmental Control Brain Computer Interface for Late-Stage ALS Patients
Taylor Connelly, Lindsay Hager, Emily Malloy, Michael Wertz, Hasan Ayaz, PhD Drexel University, Ayaz Lab
1) User Need: 1.a) Use Case Amyotrophic lateral sclerosis (ALS) is characterized by progressive degeneration of motor neurons in the spinal cord and brain, inhibiting voluntary muscle control & resulting in late-stage locked-in (LSLI) syndrome. Assistive technology is developed to allow locked-in patients to autonomously control household conditions. 1.b) Problem ImpactLack of independence impacts patient mental health, increasing frustration, anxiety, and depression. Assistive technology uses physiological activity, like ocular or motor movement, which are unsuitable for LSLI patients. Brain computer interfaces (BCI) are digital systems that detect and filter neurological signals, then classify and convert them into meaningful commands to execute decision-making tasks. Use of BCI partially returns patient autonomy. 1.c) Scope (Objective)In this application, rapid serial visual presentation (RSVP) is used to display a series of environmental control icon options. We aim to develop a non-invasive, EEG-based BCI system, with RSVP stimuli eliciting P300 responses as a selection mechanism to complete household tasks. 2) Design Inputs 2.a) Constraints: We must adhere to the universal constraints of time, budget, resources, and policy. Our solution must utilize the RSVP paradigm and elicit a P300 signal. We will use an EEG amplifier provided by Ayaz Lab. MATLAB was selected as the data processing software due to our team’s familiarity. Finally, we chose a digital assistant which is compatible with external signals. 2.b) Requirements: Our system must allow command interchangeability. Additionally, we must confirm the ability to live-stream and transmit EEG signals in real-time. The system itself must recognize the selection in 45 seconds with 70% accuracy. Finally, the selected environmental control command must be executed. 3) Solution 3.a) Design - Intended Use: See Figure 1. 3.b) Build - DEMOThe two main components are the acquisition and filtering scripts and the RSVP-UI that allows for the icon selection and visual stimuli. Both are currently partially developed, ensuring live-streaming data and icon interchangeability. 4) Verification Results4.a) Introduction Our analysis of live data acquisition, speed, accuracy and usability will then be done via a confusion matrix, considering false positives/negatives and intended choices. The system will also be tested on its interchangeability; the secondary user (caregiver) will attempt to switch out icons on the board, verifying the usability of the user interface. 4.b) MethodsRaw EEG signals were collected via MentaLab EEG Amplifier by 4 electrodes placed in the occipital and frontal lobe, while a RSVP stimuli are presented in a sequence of 100 ms per icon, to a healthy test subject.4.c) Results Live EEG signal acquisition was confirmed and displayed into MATLAB. Additionally, the interchangeability of the icons was verified via the design of an RSVP-UI application, which prompts the caregiver to select the number of icons (Figure 1). 5) Conclusion 5.a) Summary: The system has not yet been finalized, with 2 requirements met so far. We are continuing to develop the system to test speed, accuracy and command execution. 5.b) Revisions: The system would ideally be used on ALS patients to assess feasibility as an end-user solution. Additionally, the system would be transferable to non-licensed software. 5.c) Impact (Future Version): We hope that this innovative system will provide late-stage ALS patients with control over their environments and enable them to regain autonomy. References: 1. Hild et al, 2010. 10.1109/IEMBS.2010.5627081 2. Won et al, 2022, 10.1038/s41597-022-01509-w Acknowledgements: We would like to acknowledge Dr. Hasan Ayaz and his lab for their mentorship and support.