Literature provides results which show that pre-processing employing independent component analysis (ICA) to remove artefacts, can result in better accuracy (O’Sullivan et al. 2015). Improvement in accuracy has been also observed that other envelope extraction methods (Biesmans et al. 2017) that are based on auditory processing models.
Establishing accurate algorithms and methods of tracked signal identification, such as AAD, may markedly help hearing impaired persons in the nearest future, as it may be implemented in a new generation of hearing aids integrated with EEG system. Currently available literature discusses mainly variety of effects of the various degradation of the speech signal that occurs before the signal reaches the ears of the listener. Such studies explain how the loss of temporal and spectral cues in speech signal affects the neural processing. The review of these works can be found in papers by Zoefel andVan Rullen (2015) and Ding i Simon (2014). Ding and Simon (2013) have presented experimental results in which the speech was degraded by presenting it at a background noise. They have shown how energetic masking affected the neural response. In other works vocoders have been used. For example Ding et al. (2014) have used distorted speech presented in the quiet, while Ding and Simon (2013) have used a competing speech.