One does not Simply RSVP: Mental Workload to Select Speed Reading Parameters using Electroencephalography

Kosch, Thomas; Schmidt, Albrecht; Thanheiser, Simon; Chuang, Lewis L.

doi:10.1145/3313831.3376766

Cited by 11 publications

(10 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The positive results achieved using theta band EEG features might be explained by the importance of this frequency band for successful semantic processing. Theta power is expected to rise with increasing language processing activity (Kosch et al, 2020 ). Various studies have shown that theta oscillations are related to semantic memory retrieval and can be task-specific (e.g., Bastiaansen et al, 2005 ; Giraud and Poeppel, 2012 ; Marko et al, 2019 ).…”

Section: Discussionmentioning

confidence: 99%

Decoding EEG Brain Activity for Multi-Modal Natural Language Processing

Hollenstein

Renggli

Glaus

et al. 2021

Front. Hum. Neurosci.

View full text Add to dashboard Cite

Until recently, human behavioral data from reading has mainly been of interest to researchers to understand human cognition. However, these human language processing signals can also be beneficial in machine learning-based natural language processing tasks. Using EEG brain activity for this purpose is largely unexplored as of yet. In this paper, we present the first large-scale study of systematically analyzing the potential of EEG brain activity data for improving natural language processing tasks, with a special focus on which features of the signal are most beneficial. We present a multi-modal machine learning architecture that learns jointly from textual input as well as from EEG features. We find that filtering the EEG signals into frequency bands is more beneficial than using the broadband signal. Moreover, for a range of word embedding types, EEG data improves binary and ternary sentiment classification and outperforms multiple baselines. For more complex tasks such as relation detection, only the contextualized BERT embeddings outperform the baselines in our experiments, which raises the need for further research. Finally, EEG data shows to be particularly promising when limited training data is available.

show abstract

Section: Discussionmentioning

confidence: 99%

Decoding EEG Brain Activity for Multi-Modal Natural Language Processing

Hollenstein

Renggli

Glaus

et al. 2021

Front. Hum. Neurosci.

View full text Add to dashboard Cite

show abstract

“…Computer games [85], [103], [104] were included in the low-noise category. Data presentation tasks, visual [105]- [108] or auditory [109]- [111], and multiple learning tasks [13], [18], [112], [113], including code comprehension tasks [114]- [116] and mathematical tasks [117]- [120], were categorized as low-noise ones. Additionally, dual tasks with low noise were a Tetris game coupled with an oddball task [121], a Stroop task coupled with an oddball task [122], a SIMKAP task [123], and a threefold task performed on a computer and a customized AR system [14].…”

Section: A Rq (A): Classification Of Cognitive Load Experiments Based...mentioning

confidence: 99%

“…Epoching in particular was documented as an early step in preprocessing [14], [107], [109], [110], [117], [129], yet most studies applied it as a later step after basic filtering [101], [122], [138] or after filtering and BSS algorithms [93]- [96], [99], [100], [106], [112], [121], [125]- [127], [135], [139], [141]. Overlapping was rare, with an overlapping percentage of 80% [38]or 50% [105]. Regarding the epoch duration, short epochs of 1second were preferred for rapid changes in EEG data [96], [105], [112], [121], [122], [124], [138] with a time interval approximately 200ms before stimulus presentation until 800ms after stimulus presentation [14], [96], [109], [121], [122], [138].…”

Section: ) How Was Standard Preprocessing Executed In Semi-automated ...mentioning

confidence: 99%

A Comprehensive Survey of EEG Preprocessing Methods for Cognitive Load Assessment

Kyriaki,

Koukopoulos,

Fidas

2024

IEEE Access

View full text Add to dashboard Cite

Preprocessing electroencephalographic (EEG) signals during computer-mediated Cognitive Load tasks is crucial in Human-Computer Interaction (HCI). This process significantly influences subsequent EEG analysis and the efficacy of Artificial Intelligence (AI) models employed in Cognitive Load Assessment.Consequently, it stands as an indispensable procedure for developing dependable systems capable of adapting to users' cognitive capacities and constraints. We systematically analyzed fifty-seven (57) research papers on computer-mediated Cognitive Load EEG experiments published between 2018 and 2023. The preprocessing methods identified were multiple, controversial, and strongly dependent on the particularities of each experiment and the derived experimental dataset. Our investigation involved the meticulous classification of preprocessing methods based on distinct parameters, namely the degree of user intervention, the noise level, and the subject pool size. Particular attention was paid to semi-automated denoising technology since conventional methods, advanced approaches, and standardized pipelines overwhelm research, but no optimum solution is available yet. This survey is anticipated to provide a valuable contribution to the rising demand for an efficient and fully automated preprocessing approach in EEG-based computerized Cognitive Load experiments.

show abstract

“…Placebo-Control 4 [35,71,97,111] Placebo Study 12 [18,20,22,23,28,82,100,107,110,110] Acknowledge Placebo Effects as a Confound 11 [1,7,10,21,47,64,74,83,86,91,94,105] modalities. For example, one can adapt letter presentation in an E-reader to alpha oscillations; an indicator of workload in the electroencephalogram [55], reduce the number of interaction possibilities with the number of errors made by the user [45], or adapt the interface to the mood of the user [61]. Further examples include the user's context [90] and their interaction performance as well as their physiology [56] serving as implicit measurements for user interface adaptations.…”

Section: Category N Studiesmentioning

confidence: 99%

“…The availability of large-scale user data allowed recent advances in machine learning and neural networks to make robust predictions about user-related issues during interaction. This includes adaptations based on the user's perceived workload [53][54][55], emotions [5,52], and interaction difficulties due to the complex visualization of UI elements [51]. AI-based adaptive interfaces optimize usability by linking the system design to user input [43], where the overall goal of user interface adaptations is the improvement of the interaction efficiency and perceived usability [43].…”

Section: Category N Studiesmentioning

confidence: 99%

The Placebo Effect of Artificial Intelligence in Human-Computer Interaction

Kosch,

Welsch,

Chuang

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

In medicine, patients can obtain real benefits from a sham treatment. These benefits are known as the placebo effect. We report two experiments (Experiment I: N=369; Experiment II: N=100) demonstrating a placebo effect in adaptive interfaces. Participants were asked to solve word puzzles while being supported by no system or an adaptive AI interface. All participants experienced the same word puzzle difficulty and had no support from an AI throughout the experiments. Our results showed that the belief of receiving adaptive AI support increases expectations regarding the participant's own task performance, sustained after interaction.These expectations were positively correlated to performance, as indicated by the number of solved word puzzles. We integrate our findings into technological acceptance theories and discuss implications for the future assessment of AI-based user interfaces and novel technologies. We argue that system descriptions can elicit placebo effects through user expectations biasing the results of user-centered studies.CCS Concepts: • Human-centered computing → User studies; HCI theory, concepts and models; Empirical studies in HCI .

show abstract

One does not Simply RSVP: Mental Workload to Select Speed Reading Parameters using Electroencephalography

Cited by 11 publications

References 52 publications

Decoding EEG Brain Activity for Multi-Modal Natural Language Processing

Decoding EEG Brain Activity for Multi-Modal Natural Language Processing

A Comprehensive Survey of EEG Preprocessing Methods for Cognitive Load Assessment

The Placebo Effect of Artificial Intelligence in Human-Computer Interaction

Contact Info

Product

Resources

About