Automated Loudness Growth Prediction From EEG Signals Using Autoencoder and Multi-Target Regression

Accurately assessing loudness perception is crucial for optimizing hearing aid fittings, particularly for those unable to perform subjective tests. This study presents an automated method for estimating frequency-specific loudness growth curves using tone-burst auditory brainstem responses (ABRs), which are a subset of EEG (electroencephalography) signals. Unlike traditional methods that rely on manually engineered features, the proposed method uses convolutional autoencoder to learn latent representations of ABR signals, reducing dimensionality while preserving critical auditory information. The extracted features are mapped to psychoacoustic loudness growth estimates using a multi-target regression model based on a convolutional neural network. We conduct an ablation study to analyze the impact of different autoencoder configurations on feature extraction performance. The results demonstrate strong predictive consistency, with high Pearson correlation coefficients (PCC ≥ 0.9) and low mean square errors (MSE ≤ 0.0011) across different stimulus frequencies and subjects. The implementation code will be shared on GitHub after the paper is accepted.