A Brute-Force CNN Model Selection for Accurate Classification of Sensorimotor Rhythms in BCIs

Abstract

The ultimate goal of Brain-Computer Interface (BCI) research is to enable individuals to interact with their environment by translating their mental imagery. In this regard, a salient issue is the identification of brain activity patterns that can be used to classify intention. Using Electroencephalographic (EEG) signals as archetypical, this classification problem generally possesses two stages: (i) extracting features from collected EEG waveforms; and (ii) constructing a classifier using extracted features. With the advent of deep learning, however, the former stage is generally absorbed into the latter. Nevertheless, the burden has now shifted from trying a number of feature extraction methods to tuning a large number of hyperparameters and architectures. Among existing deep learning architectures used in BCI, Convolutional Neural Networks (CNN) have become an attractive choice. Most of the existing studies that use these networks are based on well-known architectures such as AlexNet or ResNet, use the domain knowledge to construct the final architecture or have an unclear strategy deployed for model selection. This raises the question as to whether constructing accurate CNN-based classifiers is possible using a principled model selection, with the most straightforward one being the brute-force search or, alternatively, experience and developing high intuition regarding hyperparameters combined with an ad hoc approach is the most prudent way to go about designing them. To this end, in this paper, we first define a space of hyperparameters restricted by our computing power. Then we show that an exhaustive search within this limited space of CNN hyperparameters leads to accurate classification of sensorimotor rhythms that arise during motor imagery tasks.

Investigation of a Deep-Learning Based Brain–Computer Interface With Respect to a Continuous Control Application

“…For the classification task at hand, there are already several CNN models from other scientific works [13], [14], [6]. Essential for the intended BCI application, the chosen EEGNet [3] has been developed with low power embedded system applications in mind.…”

Spatio-temporal transformers for decoding neural movement control

“…On the other hand, the broad spectrum of recording technologies and the resulting abundance of experimental data have posed challenges in developing standardized routines of analysis consistent across species and conditions. In such a plethora of information, machine learning algorithms, especially deep learning ones, are emerging as cutting-edge tools for skimming redundancies and selecting the most relevant variables for complex system's dynamics [20][21][22][23][24][25], opening exciting opportunities for analysis and interpretation of such data [26,27]. In particular, their application in cognitive neuroscience has gained significant attention for their appealing flexibility to link neural representations to behavioral outcomes [28].…”

Data Constraints and Performance Optimization for Transformer-Based Models in EEG-Based Brain-Computer Interfaces: A Survey

“…These applications extend from providing lifelines of communication for individuals struggling with debilitating conditions such as amyotrophic lateral sclerosis (ALS) [1] and locked-in syndrome [2], to the identification and manage-ment of epileptic seizures [3]. Furthermore, the integration of EEG-based BCIs into cutting-edge prosthetics [4], immersive gaming environments [5], virtual reality platforms [6], and the broader domain of scientific research underscores the transformative impact of this technology [7].…”