This paper considers the auditory attention detection (AAD) paradigm, where the goal is to determine which of two simultaneous speakers a person is attending to. The paradigm relies on recordings of the listener's brain activity, e.g., from electroencephalography (EEG). To perform AAD, decoded EEG signals are typically correlated with the temporal envelopes of the speech signals of the separate speakers. In this paper, we study how the inclusion of various degrees of auditory modelling in this speech envelope extraction process affects the AAD performance, where the best performance is found for an auditory-inspired linear filter bank followed by power law compression. These two modelling stages are computationally cheap, which is important for implementation in wearable devices, such as future neuro-steered auditory prostheses. We also introduce a more natural way to combine recordings (over trials and subjects) to train the decoder, which reduces the dependence of the algorithm on regularization parameters. Finally, we investigate the simultaneous design of the EEG decoder and the audio subband envelope recombination weights vector using either a norm-constrained least squares or a canonical correlation analysis, but conclude that this increases computational complexity without improving AAD performance.

Speech intelligibility is currently measured by scoring how well a person can identify a speech signal. The results of such behavioral measures reflect neural processing of the speech signal, but are also influenced by language processing, motivation, and memory. Very often, electrophysiological measures of hearing give insight in the neural processing of sound. However, in most methods, non-speech stimuli are used, making it hard to relate the results to behavioral measures of speech intelligibility. The use of natural running speech as a stimulus in electrophysiological measures of hearing is a paradigm shift which allows to bridge the gap between behavioral and electrophysiological measures. Here, by decoding the speech envelope from the electroencephalogram, and correlating it with the stimulus envelope, we demonstrate an electrophysiological measure of neural processing of running speech. We show that behaviorally measured speech intelligibility is strongly correlated with our electrophysiological measure. Our results pave the way towards an objective and automatic way of assessing neural processing of speech presented through auditory prostheses, reducing confounds such as attention and cognitive capabilities. We anticipate that our electrophysiological measure will allow better differential diagnosis of the auditory system, and will allow the development of closed-loop auditory prostheses that automatically adapt to individual users.

Current EEG artifact removal techniques often have limited applicability due to their specificity to one kind of artifact, their complexity, or simply because they are too 'blind'. This paper demonstrates a fast, robust and generic algorithm for removal of EEG artifacts of various types, i.e. those that were annotated as unwanted by the user.

APEX 3 is a software test platform for auditory behavioral experiments. It provides a generic means of setting up experiments without any programming. The supported output devices include sound cards and cochlear implants from Cochlear Corporation and Advanced Bionics Corporation. Many psychophysical procedures are provided and there is an interface to add custom procedures. Plug-in interfaces are provided for data filters and external controllers. APEX 3 is supported under Linux and Windows and is available free of charge.

When we grow older, understanding speech in noise becomes more challenging. Research has demonstrated the role of auditory temporal and cognitive deficits in these age-related speech-in-noise difficulties. To better understand the underlying neural mechanisms, we recruited young, middle-aged, and older normal-hearing adults and investigated the interplay between speech understanding, cognition, and neural tracking of the speech envelope using electroencephalography. The stimuli consisted of natural speech masked by speech-weighted noise or a competing talker and were presented at several subject-specific speech understanding levels. In addition to running speech, we recorded auditory steady-state responses at low modulation frequencies to assess the effect of age on nonspeech sounds. The results show that healthy aging resulted in a supralinear increase in the speech reception threshold, i.e., worse speech understanding, most pronounced for the competing talker. Similarly, advancing age was associated with a supralinear increase in envelope tracking, with a pronounced enhancement for older adults. Additionally, envelope tracking was found to increase with speech understanding, most apparent for older adults. Because we found that worse cognitive scores were associated with enhanced envelope tracking, our results support the hypothesis that enhanced envelope tracking in older adults is the result of a higher activation of brain regions for processing speech, compared with younger adults. From a cognitive perspective, this could reflect the inefficient use of cognitive resources, often observed in behavioral studies. Interestingly, the opposite effect of age was found for auditory steady-state responses, suggesting a complex interplay of different neural mechanisms with advancing age. NEW & NOTEWORTHY We measured neural tracking of the speech envelope across the adult lifespan and found a supralinear increase in envelope tracking with age. Using a more ecologically valid approach than auditory steady-state responses, we found that young and older, as well as middle-aged, normal-hearing adults showed an increase in envelope tracking with increasing speech understanding and that this association is stronger for older adults.

Current research on AAD always assumes the availability of the clean speech signals, which limits the applicability in real settings. We have extended this research to detect the attended speaker even when only microphone recordings with noisy speech mixtures are available. This is an enabling ingredient for new brain-computer interfaces and effective filtering schemes in neuro-steered hearing prostheses. Here, we provide a first proof of concept for EEG-informed attended speaker extraction and denoising.

11Speech intelligibility is currently measured by scoring how well a person can identify 12 a speech signal. The results of such behavioral measures reflect neural processing of the 13 speech signal, but are also influenced by language processing, motivation and memory. Very 14 often electrophysiological measures of hearing give insight in the neural processing of sound. 15However, in most methods non-speech stimuli are used, making it hard to relate the re- 1. CC-BY-NC-ND 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint . http://dx.doi.org/10.1101/246660 doi: bioRxiv preprint first posted online bridge the gap between behavioral and electrophysiological measures. Here, by decoding

When listening to speech, our brain responses time lock to acoustic events in the stimulus. Recent studies have also reported that cortical responses track linguistic representations of speech. However, tracking of these representations is often described without controlling for acoustic properties. Therefore, the response to these linguistic representations might reflect unaccounted acoustic processing rather than language processing. Here, we evaluated the potential of several recently proposed linguistic representations as neural markers of speech comprehension. To do so, we investigated EEG responses to audiobook speech of 29 participants (22 females). We examined whether these representations contribute unique information over and beyond acoustic neural tracking and each other. Indeed, not all of these linguistic representations were significantly tracked after controlling for acoustic properties. However, phoneme surprisal, cohort entropy, word surprisal, and word frequency were all significantly tracked over and beyond acoustic properties. We also tested the generality of the associated responses by training on one story and testing on another. In general, the linguistic representations are tracked similarly across different stories spoken by different readers. These results suggests that these representations characterize the processing of the linguistic content of speech. Significance StatementFor clinical applications, it would be desirable to develop a neural marker of speech comprehension derived from neural responses to continuous speech. Such a measure would allow for behavior-free evaluation of speech understanding; this would open doors toward better quantification of speech understanding in populations from whom obtaining behavioral measures may be difficult, such as young children or people with cognitive impairments, to allow better targeted interventions and better fitting of hearing devices.