Robust Subject-Independent P300 Waveform Classification via Signal Pre-Processing and Deep Learning

Abstract

Brain Computer Interfaces (BCIs) are capable of processing neural stimuli using electroencephalogram (EEG) measurements to aid communication capabilities. Yet, BCIs often require extensive calibration steps in order to be tuned to specific users. In this work, we develop a subject independent P300 classification framework, which eliminates the need for user-specific calibration. We begin by employing a series of pre-processing steps, where, among other steps, we consider different trial averaging methodologies and various EEG electrode configurations. We then consider three distinct deep learning architectures and two linear machine learning models as P300 signal classifiers. Through evaluation on three datasets, and in comparison to three benchmark P300 classification frameworks, we find that averaging up to seven trials while using eight specific electrode channels on a two-layered convolutional neural network (CNN) leads to robust subject independent P300 classification. In this capacity, our method achieves greater than a 0.20 gain in AUC in comparison to prior P300 classification methods. In addition, our proposed framework is computationally efficient with training time gains of greater than 3x, compared to linear machine learning models, and online evaluation time speedups of up to 2x compared to benchmark methods.

Signal alignment for cross-datasets in P300 brain-computer interfaces

“…However, given the large number of features, conventional machine learning techniques, such as LDA and SVM, may not be optimal choices, necessitating careful feature selection. Deep learning models, as demonstrated in prior studies [13,48], offer promise in both feature selection and classification tasks. Thus, investigating the synergies between SA and classification methods (xDAWN [9], Riemannian geometry [49], and deep learning models) is an interesting topic.…”

Challenges of brain-computer interface facilitated cognitive assessment for children with cerebral palsy

“…While calibration data of this duration is not necessarily essential, calibrations performed on small amounts of data may report success, but not correctly interpret new data. Methods to create BCIs that work without individualized calibration or that perform more rapid calibration using transfer learning from past participants are not yet readily available [e.g., ( Lu et al, 2009 ; Xu et al, 2015 ; Sahay and Brinton, 2021 )]. Methods are needed to automatically remove data with artifact or low participant attention.…”