Dr. Günet Eroğlu is a researcher and entrepreneur in the fields of neurotechnology, artificial intelligence, and clinical decision-support systems. With an academic background in biostatistics and applied AI, her work focuses on EEG-based neurophysiological analysis, adaptive neurofeedback systems, and AI-driven personalization in developmental disorders such as ADHD and dyslexia. She is the founder of Auto Train Brain and NeuroSphere, platforms that integrate artificial intelligence with neurofeedback technologies to enhance evidence-based clinical practice. Her research combines reinforcement learning, signal processing, and clinical outcome modeling to develop scalable, patient-centered digital health solutions. Dr. Eroğlu has authored multiple peer-reviewed publications on EEG biomarkers, neuroinflammation, and AI-assisted neurodevelopmental screening. Her mission is to bridge advanced computational methods with practical healthcare applications to improve global patient care and accessibility in neurodevelopmental services.

Explainable EEG Biomarkers for Personalized Neuroadaptive Systems: Bridging Artificial Intelligence and Neurofeedback The integration of artificial intelligence (AI) into neurofeedback systems has significantly expanded the capacity to analyze complex electroencephalography (EEG) signals and deliver adaptive, individualized interventions. However, many AI-driven neurofeedback solutions remain limited by their reliance on opaque, black-box models, restricting clinical interpretability and trust. This study addresses this gap by proposing an explainable artificial intelligence (XAI) framework for identifying EEG-based biomarkers that support transparent and personalized neuroadaptive systems. Resting-state and task-related EEG data were analyzed using supervised machine learning models to extract multidimensional neural features, including spectral profile morphology, frequency-band interactions, and functional connectivity patterns across cortical regions. To enhance interpretability, SHapley Additive exPlanations (SHAP) were employed to quantify feature-level contributions to model predictions and to track longitudinal changes associated with neurofeedback training. This approach enables the identification of individualized neural signatures that reflect learning-related cortical adaptation rather than group-level averages. The results demonstrate that explainable EEG biomarkers provide clinically meaningful insights into neural dynamics underlying attention, cognitive control, and learning processes. Importantly, the XAI-driven framework allows neurofeedback protocols to be dynamically adjusted based on subject-specific EEG responses, supporting truly personalized neuroadaptive interventions. By bridging AI-driven analytics with neurophysiological interpretability, this work advances the development of transparent, data-driven neurofeedback systems and contributes to the broader goal of precision neurotechnology in both clinical and performance-oriented applications.