Iustina Rotaru

Iustina Rotaru

Auto-generated profile, based on publicly available information

Iustina Rotaru

Department of Neurosciences, ExpORL, KU Leuven, 3000 Leuven, Belgium

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 February 2024 in Journal of Neural Engineering
Verified
Journal Article
Closed access
IRIustina Rotaru
Iustina Rotaru
SGSimon Geirnaert
Simon Geirnaert
NHNicolas Heintz
Nicolas Heintz
IRIris Van de Ryck
Iris Van de Ryck
ABAlexander Bertrand
Alexander Bertrand
TFTom Francart
Tom Francart
Objective. Spatial auditory attention decoding (Sp-AAD) refers to the task of identifying the direction of the speaker to which a person is attending in a multi-talker setting, based on the listener’s neural recordings, e.g. electroencephalography (EEG). The goal of this study is to thoroughly investigate potential biases when training such Sp-AAD decoders on EEG data, particularly eye-gaze biases and latent trial-dependent confounds, which may result in Sp-AAD models that decode eye-gaze or trial-specific fingerprints rather than spatial auditory attention. Approach. We designed a two-speaker audiovisual Sp-AAD protocol in which the spatial auditory and visual attention were enforced to be either congruent or incongruent, and we recorded EEG data from sixteen participants undergoing several trials recorded at distinct timepoints. We trained a simple linear model for Sp-AAD based on common spatial patterns filters in combination with either linear discriminant analysis (LDA) or k-means clustering, and evaluated them both across- and within-trial. Main results. We found that even a simple linear Sp-AAD model is susceptible to overfitting to confounding signal patterns such as eye-gaze and trial fingerprints (e.g. due to feature shifts across trials), resulting in artificially high decoding accuracies. Furthermore, we found that changes in the EEG signal statistics across trials deteriorate the trial generalization of the classifier, even when the latter is retrained on the test trial with an unsupervised algorithm. Significance. Collectively, our findings confirm that there exist subtle biases and confounds that can strongly interfere with the decoding of spatial auditory attention from EEG. It is expected that more complicated non-linear models based on deep neural networks, which are often used for Sp-AAD, are even more vulnerable to such biases. Future work should perform experiments and model evaluations that avoid and/or control for such biases in Sp-AAD tasks.

11 January 2026 in Hearing Research
Verified
Journal Article
Closed access
IRIris Van de Ryck
Iris Van de Ryck
NHNicolas Heintz
Nicolas Heintz
IRIustina Rotaru
Iustina Rotaru
SGSimon Geirnaert
Simon Geirnaert
ABAlexander Bertrand
Alexander Bertrand
TFTom Francart
Tom Francart

4 October 2025 in Hearing Research
Verified
Journal Article
Closed access
IRIris Van de Ryck
Iris Van de Ryck
NHNicolas Heintz
Nicolas Heintz
IRIustina Rotaru
Iustina Rotaru
DFDebora Fieberg
Debora Fieberg
ABAlexander Bertrand
Alexander Bertrand
TFTom Francart
Tom Francart

27 May 2025 in bioRxiv
Verified
Preprint
Open access
PDF file
IRIris Van de Ryck
Iris Van de Ryck
NHNicolas Heintz
Nicolas Heintz
IRIustina Rotaru
Iustina Rotaru
DFDebora Fieberg
Debora Fieberg
ABAlexander Bertrand
Alexander Bertrand
TFTom Francart
Tom Francart
Objectives Auditory Attention Decoding (AAD) is a technique utilizing brain signals to decode on which sound the listener focuses the attention. In most current studies, the effect of type of speech materials used and sex of the listener is not considered. We investigated the effect on AAD performance of factors related to the speaker (such as the sex of the speaker, background noise level, and same versus mixed-sex conditions) and the listener (sex of the listener). Design Forty-two young adults with normal hearing participated in the study. They listened to 2 competing speakers and were instructed to attend to one speaker and ignore the other speaker, whilst electroencephalography (EEG) and electrooculography (EOG) were measured. Background noise was introduced in half of the conditions. AAD performance was compared across eight experimental conditions. Results A significant main effect of speaker sex was found: A male target and/or male masker speaker resulted in higher AAD performance compared to a female speaker with a higher fundamental frequency. These effects were found to be small and therefore likely clinically irrelevant. Conclusion While no substantial effects were found on the factors investigated in this study, including diverse and realistic training scenarios remains a valuable approach to prevent potential influences from other factors.

22 June 2025 in bioRxiv
Verified
Preprint
Open access
PDF file
IRIris Van de Ryck
Iris Van de Ryck
NHNicolas Heintz
Nicolas Heintz
IRIustina Rotaru
Iustina Rotaru
SGSimon Geirnaert
Simon Geirnaert
ABAlexander Bertrand
Alexander Bertrand
TFTom Francart
Tom Francart
Objectives Auditory attention decoding (AAD) refers to the process of identifying which sound source a listener is attending to, based on neural recordings, such as electroencephalography (EEG). Most AAD studies use a competing speaker paradigm where two continuously active speech signals are simultaneously presented, in which the participant is instructed to attend to one speaker and ignore the other speaker. However, such a competing two-speaker scenario is uncommon in real life, as speakers typically take turns rather than speaking simultaneously. In this paper, we argue that decoding attention to conversations (rather than individual speakers) is a more relevant paradigm for testing AAD algorithms. In such a conversation-tracking paradigm, the AAD algorithm focusses on switching between entire conversations, resulting in less frequent attention shifts (ignoring turn-taking within conversations), thereby allowing for more relaxed constraints on the decision time. Design To test AAD performance in such a conversation-tracking paradigm, we simulated a challenging restaurant scenario with three simultaneous two-speaker conversations, which were podcasts presented in front of the listener and in the back left and back right of the room. We conducted an EEG experiment on 20 normal-hearing participants to compare the performance of AAD in the commonly used competing speaker paradigm with two speakers versus the conversation tracking paradigm with 2 or 3 conversations, each containing two turn-taking speakers. Results We found that AAD, using stimulus decoding, worked well under all experimental conditions, and that the accuracy was not influenced by the direction of attention, the proximity to the target conversation, or the presence of within-trial attention switches (versus a condition with sustained attention). Given the challenging scenario, we probed for the participants’ listening experience and found a correlation between the neural decoding performance and the perceived listening effort and self-reported speech intelligibility. To gain insight into the speech intelligibility of the participants in our setup, they performed a speech-in-noise test (Flemish matrix sentence test), but we did not find a correlation between the speech intelligibility performance and the AAD performance.

15 July 2023 in bioRxiv
Verified
Preprint
Open access
PDF file
IRIustina Rotaru
Iustina Rotaru
SGSimon Geirnaert
Simon Geirnaert
NHNicolas Heintz
Nicolas Heintz
IRIris Van de Ryck
Iris Van de Ryck
ABAlexander Bertrand
Alexander Bertrand
TFTom Francart
Tom Francart
Objective: Auditory attention decoding (AAD) refers to the task of identifying which speaker a person is listening to in a multi-talker setting, based on their neural recordings. The Common Spatial Patterns (CSP) algorithm has previously shown promising potential w.r.t. the state-of-the-art AAD algorithms to create discriminative features from electroencephalography (EEG) signals in a task of spatial AAD (sAAD). However, there has been some skepticism related to the underlying decoding mechanisms of such sAAD methods, as well as their generalization capabilities across subjects and experimental trials. In this study, we aimed to investigate (1) what type of mechanisms (neural vs. non-neural) drive the CSP decoding and (2) how well CSP filters can generalize across trials and subjects. Approach: We designed a two-speaker audiovisual sAAD protocol in which we enforced the spatial auditory and visual attention to be either congruent or incongruent, and we recorded EEG data from sixteen participants performing this task. Main results: Firstly, we found that the sAAD accuracy with CSP-derived features was significantly higher in scenarios where the target visual and auditory stimuli were co-located, potentially indicating that CSP decoders exploited eye-gaze information. Secondly, CSP decoding remained feasible even without relevant eye-gaze information, i.e., when the location of the attended visual target was continuously shifted to ensure spatial dissociation with the auditory stimulus' location. This finding suggests that CSPs are able to extract neural lateralization patterns reflecting spatial auditory attention independent of the eye-gaze direction. Thirdly, we identified a limitation in the between-trial generalization ability of the CSP feature embeddings, observing strong distribution shifts in the feature space across trials. However, we demonstrated this can be overcome by employing partially-unsupervised classification methods. Significance: Collectively, our findings confirm the feasibility of CSP filters in decoding the locus of auditory attention in various AV conditions, while equally emphasizing the need for novel algorithms that are robust to generalization.