Publications

A critical study on the robustness of BCI models, proving that Transformers learn true neural intent while CNNs often rely on ocular or muscular artifacts.

An investigation into how CNNs often learn artifacts instead of true neural intent, and how Transformer models provide a more robust solution for rehabilitation BCI.

Official Persian translation of the seminal BCI research series (Volume 12). This work bridges neuroscience and AI, making advanced BCI knowledge accessible to the Persian-speaking academic community.

This paper presents an automated approach using a Vision Transformer (ViT) model to improve the accuracy of bone age prediction from hand X-ray images, achieving a Mean Absolute Error of 3.2 months.

This research presents a novel Transformer architecture to decode imagined speech from EEG signals, achieving significant improvements over baseline CNN models.