Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a pipeline for EEG data classification using synthetic data augmentation, LSTM networks, and fine-tuning. Synthetic EEG data, generated to address data scarcity, is used to pre-train an LSTM, which is then fine-tuned on real EEG data. The results demonstrate improved classification performance on implicit visual stimuli experiments, specifically the Necker cube.
Synthetic data + LSTM fine-tuning unlocks improved classification of raw EEG data, even with limited real-world examples.
In this paper, we discuss a Machine Learning pipeline for the classification of EEG data. We propose a combination of synthetic data generation, long short-term memory artificial neural network (LSTM), and fine-tuning to solve classification problems for experiments with implicit visual stimuli, such as the Necker cube with different levels of ambiguity. The developed approach increased the quality of the classification model of raw EEG data.