Interpreting cognitive state is essential for understanding children's learning abilities, including attention, memory, and problem- solving skills, which directly affect academic and social development. Early identification of cognitive irregularities enables timely assessment and individualized support plans. Electroencephalography (EEG) is a non- invasive, effective method for recording neuromorphic brain signals associated with cognitive processes. This study presents a novel hybrid optimization algorithm inspired by the biological visual system to interpret cognitive state from EEG signals recorded by an open- source, low- cost, portable bio- neuromorphic sensing platform (OpenBCI). A robust preprocessing and discriminative feature- extraction method is used to model the complex spatiotemporal patterns of EEG with a Hybrid Convolutional Neural Network (HCNN). To improve classification effectiveness and convergence stability, a Hybrid Optimization (HO) strategy is presented that combines Particle Swarm Optimization (PSO), Grasshopper and Grey Wolf Optimization (G2WO). The hybrid PSO–GWO method leverages collective swarm intelligence and hierarchical hunting behaviour to optimize the learning parameters of a CNN. Results show that the proposed HO- optimized CNN model successfully differentiates between cognitive states with 92. 4% accuracy when applied to OpenBCI EEG signals. Longer chains yield better performance, with extended HO–PSO–G2WO achieving 95.2% accuracy, 93.1% precision and 92.2% recall at the optimal learning rate of 0. 02. Similarly, the Hybrid Optimization on Grasshopper & Grey scale Optimization method (Hybrid PSO + G2WO) offers improvement with 98. 98.21% accuracy, 98. 2% recall, and 98. 98.18% precision at an execution time of 13. 5 s for the provided pre-processed dataset of cognitive- based EEG signals. This research effort is directed toward the reliable classification of cognitive state, not directly to the enhancement of cognitive skills, since it was not tested against any intervention. However, the results demonstrate the efficacy of bio-inspired hybrid optimization in improving DL performance for EEG-based cognitive interpretation. The proposed framework lays a solid foundation for future longitudinal studies and personalized cognitive assessment systems from a neuromorphic EEG perspective.