Toward Improving fNIRS Classification: A Study on Activation Functions in Deep Neural Architectures

Activation functions are critical to the performance of deep neural networks, particularly in domains such as functional near-infrared spectroscopy (fNIRS), where nonlinearity, low signal-to-noise ratio (SNR), and signal variability poses significant challenges to model accuracy. However, the impact of activation functions on deep learning (DL) performance in the fNIRS domain remains underexplored and lacks systematic investigation in the current literature. This study evaluates a range of conventional and field-specific activation functions for fNIRS classification tasks using multiple deep learning architectures, including the domain-specific fNIRSNet, AbsoluteNet, MDNN, and shallowConvNet (as the baseline), all tested on a single dataset recorded during an auditory task. To ensure fair a comparison, all networks were trained and tested using standardized preprocessing and consistent training parameters. The results show that symmetrical activation functions such as Tanh and the Absolute value function Abs(x) can outperform commonly used functions like the Rectified Linear Unit (ReLU), depending on the architecture. Additionally, a focused analysis of the role of symmetry was conducted using a Modified Absolute Function (MAF), with results further supporting the effectiveness of symmetrical activation functions on performance gains. These findings underscore the importance of selecting proper activation functions that align with the signal characteristics of fNIRS data.

Precision Enhancement in Sustained Visual Attention Training Platforms: Offline EEG Signal Analysis for Classifier Fine-Tuning

In this study, a novel open-source brain-computer interface (BCI) platform was developed to decode scalp electroencephalography (EEG) signals associated with sustained attention. The EEG signal collection was conducted using a wireless headset during a sustained visual attention task, where participants were instructed to discriminate between composite images superimposed with scenes and faces, responding only to the relevant subcategory while ignoring the irrelevant ones. Seven volunteers participated in this experiment. The data collected were subjected to analyses through event-related potential (ERP), Hilbert Transform, and Wavelet Transform to extract temporal and spectral features. For each participant, utilizing its extracted features, personalized Support Vector Machine (SVM) and Random Forest (RF) models with tuned hyperparameters were developed. The models aimed to decode the participant's attentional state towards the face and scene stimuli. The SVM models achieved a higher average accuracy of 80\% and an Area Under the Curve (AUC) of 0.86, while the RF models showed an average accuracy of 78\% and AUC of 0.8. This work suggests potential applications for the evaluation of visual attention and the development of closed-loop brainwave regulation systems in the future.