Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Device Technologies and Biomedical Robotics

Device Technologies and Biomedical Robotics - Poster Session D

Poster I15 - Optimizing the Spatial and Temporal Features of Evoked Potentials in a Low-Cost Brain Computer Interface

Friday, October 25, 2024

3:30 PM - 4:30 PM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

Hunter I. Hayes (he/him/his)

Student Researcher
California State University, Long Beach
chino, California, United States
Dan Le

Undergraduate Student Researcher
California State University, Long Beach, California, United States

Co-Author(s)

JL

Jalen Leyretana

Undergraduate Student Researcher
California State University, Long Beach, California, United States
TL

Turath Liebenau

Undergraduate Student Researcher
California State University, Long Beach, United States

Primary Investigator(s)

Derek Hu, PhD (he/him/his)

Assistant Professor
California State University, Long Beach
Long Beach, California, United States

Introduction: Brain-Computer Interfaces (BCI) provide a direct interaction between the brain’s electrical activity and an external device, assisting people with motor disabilities such as amyotrophic lateral sclerosis. BCIs often utilize event-related potential responses as an input signal; one such signal is the steady state visual evoked potential (SSVEP), which is used for its great signal to noise ratios. Despite the potential applications of BCI for patient-use at home, they are limited by high-cost hardware and challenges in data collection and analysis. Here, we aim to improve the signals to noise ratio of a low-cost SSVEP-based BCI setups by modifying the spatial and temporal parameters for human-computer interactions.

Materials and

Methods: EEG data were recorded from three healthy male subjects between 21 and 26 years of age using sixteen electrodes (F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T3, T4, T5, T6, Cz, and Pz) at a sampling rate of 125 Hz. Each subject was tasked to look at a blank screen (relaxed state), followed by a 30 second flashing stimulus (target), and then a short rest. This process was repeated for 12 trials and for 13 Hz, 15 Hz, and 17 Hz. The EEG data were demeaned, bandpass filtered from 8-22 Hz, and transformed into frequency domain using Fast Fourier Transform. SSVEP strength was measured as the ratio of power between the target flashing stimulus to the non-target frequencies. For each trial, we then selected which channels and epoch lengths yielded significantly greater SSVEP strength, optimizing parameters for future use.

Results, Conclusions, and Discussions: The mean SSVEP strength of all trials using sixteen channels and 30-second epochs was 1.1, indicating that the power of the target stimulus was greater than the power of non-target frequencies at each stimulus frequency. Channels F3 (SSVEP strength of 4.74), O1 (1.5), T3 (1.34), C4 (1.21), and O2 (1.17) had significantly greater SSVEP strength compared to other channels, suggesting that central occipital channels typically yielded the greatest signal response. The SSVEP strength increased in time with longer epoch lengths, stabilizing after ten seconds, indicating that ten second trials would be sufficient for capturing SSVEP signals. After adjusting the channel inclusion and epoch length, the mean SSVEP strength significantly improved from an average band power ratio of 1.1 to 2.56.

Our results show that SSVEPs can be reliably used as an input signal for low-cost EEG setups. Optimization of the setup through channel selection significantly increased the signal-to-noise ratio of the target signal, leading to clearer detections. Furthermore, testing the stability of the SSVEP strength over different epoch lengths showed that only ten seconds of data for each trial was needed to capture clear SSVEP signals. Future work on this would further validate these findings, as we would test these new parameters in a blinded study on a greater cohort size. By optimizing these parameters, we have demonstrated the potential for developing affordable BCIs for home use using a simple setup procedure.

Acknowledgements (Optional):