Abstract: We use an old-new recognition memory task to investigate the correlates of high and low decision confidence throughout all stages of the memory process. Group-level ERP analysis and single-trial and single-subject classification are performed on four stages of the task (information encoding, retrieval, old/new decision formation, and evaluative feedback processing). The study shows that decision confidence is significantly reflected on a group, as well as on a single-trial basis, in all investigated stages at the neural level, except during encoding. The most pronounced differences between high and low confidence can be found in the ERPs during feedback presentation after a correct answer, whereas almost no differences can be found following a wrong answers. In the feedback stage, the two levels of confidence can be separated with classification accuracies of up to 70 % on average, therefore showing potential to be used as a control state in a BCI application.

@article{krumpe2020decision, author = {Krumpe, Tanja and Gerjets, Peter and Rosenstiel, Wolfgang and Spüler, Martin}, year = {2020}, month = {jan}, pages = {1–16}, title = {Decision confidence: EEG correlates of confidence in different phases of an old/ new recognition task}, journal = {Brain-Computer Interfaces}, doi = {10.1080/2326263X.2019.1708539}, month_numeric = {1} }

Abstract: This dissertation evaluates how methodologies from machine learning can be applied in experimental psychology to gain new insights from neurophysiological data. Using two examples from memory psychology, the experimentally collected EEG data are evaluated once with classical group-level statistics and once with classification methods from the field of machine learning. The combination of the results of both methods shows that new insights can be gained that will profitably advance research in experimental psychology. The use of new methodologies in this area is necessary because conventional group-level statistics have problems that have spread extensively in science and have had serious consequences, especially in the replication crisis that started in the year 2000. The benefits of machine learning can help to alleviate these problems. In comparison to the use of group-level statistics alone, the combination of both methods allows data to be evaluated equally at both group and single-subject levels in order to obtain a complete picture of the data. Also, the information of individual regions can be compared and evaluated with that of an entire association of sensors collecting data. In this way, underlying patterns can also be considered. The addition of machine learning also enables explorative data analysis, which is not yet feasible in the area of group statistics. In concrete terms, the application of machine learning techniques has made it possible to refine the characterization of executive functions and to draw up new hypotheses regarding episodic memory. Of great importance were methodologies that make the operation of machine learning processes transparent. This allowed the application to be legitimized and the results to be interpreted for a specific purpose. Furthermore, the comparison of the behavioral accuracy and the accuracy of the machine learning process was particularly valuable. In this comparison it could be shown that there is not necessarily a connection between the visual processing of an image and its active recognition. Both case studies were able to show in a representative manner which possibilities arise from the use of machine learning methods and thus present new findings which would not have been possible without the application of machine learning in this context.

@phdthesis{krumpe2020using, author = {Krumpe, Tanja}, title = {Using machine learning as a research tool in experimental psychology}, school = {University of Tübingen}, year = {2020}, month = {jan}, doi = {10.15496/publikation-38278}, month_numeric = {1} }

Abstract: A simple recognition task was used to investigate if the item familiarity of pictures can be predicted based on single trial ERPs during item presentation, to explore the possibility of using this property in a BCI application. Two experimental parts with equal learning phases but different ratios of old and new stimuli in a forced choice memory recognition test have been performed. We were able to predict item familiarity with accuracies above 70 % based on the ERPs elicited during item representation in both parts of the experiment. In some cases, the classification accuracy even exceeds the behavioral accuracy of the subjects. Usage of this property, for example in an education-oriented scenario, seems feasible in a BCI application.

@inproceedings{TWM022019, author = {Krumpe, Tanja and Rosenstiel, Wolfgang and Spüler, Martin}, title = {Prediction of item familiarity based on ERPs }, booktitle = {7th International Conference on Brain-Computer Interface (BCI), South Korea}, year = {2019}, month = {feb}, organization = {IEEE}, doi = {10.1109/IWW-BCI.2019.8737330}, month_numeric = {2} }

Abstract: We use an old-new recognition memory task to investigate the correlates of high and low decision confidence throughout all stages of the memory process. Group-level ERP analysis and single-trial and single-subject classification are performed on four stages of the task (information encoding, retrieval, old/new decision formation, and evaluative feedback processing). The study shows that decision confidence is significantly reflected on a group, as well as on a single-trial basis, in all investigated stages at the neural level, except during encoding. The most pronounced differences between high and low confidence can be found in the ERPs during feedback presentation after a correct answer, whereas almost no differences can be found following a wrong answers. In the feedback stage, the two levels of confidence can be separated with classification accuracies of up to 70 % on average, therefore showing potential to be used as a control state in a BCI application.

@inproceedings{TPWM052018, author = {Krumpe, Tanja and Gerjets, Peter and Rosenstiel, Wolfgang and Spüler, Martin}, title = {Decision confidence: EEG correlates of confidence in different phases of an old/new recognition task}, booktitle = {Proceedings of the 7th International BCI Meeting}, year = {2018}, month = {may}, pages = {180-181}, address = {Asilomar, CA}, doi = {10.1080/2326263X.2019.1708539}, month_numeric = {5} }

@inproceedings{TKWM092017, author = {Krumpe, Tanja and Baumgärtner, K. and Rosenstiel, Wolfgang and Spüler, Martin}, title = {Non-stationarity and inter-subject variability of EEG characteristics in the context of BCI development}, booktitle = {Proceedings of the 7th Graz Brain-Computer Interface Conference}, year = {2017}, month = {sep}, pages = {260-265}, month_numeric = {9} }

@article{MTCCWP2017, author = {Spüler, Martin and Krumpe, Tanja and Walter, Carina and Scharinger, Christian and Rosenstiel, Wolfgang and Gerjets, Peter}, title = {Brain-computer interfaces for educational applications}, journal = {Informational Environments : Effects of Use, Effective Designs}, year = {2017}, pages = {177-201} }

@inproceedings{TCPWM062016, author = {Krumpe, Tanja and Scharinger, Christian and Gerjets, Peter and Rosenstiel, Wolfgang and Spüler, Martin}, title = {Disentangeling working memory load - finding inhibition and updating components in EEG data}, booktitle = {Proceedings of the 6th International Brain-Computer Interface Meeting}, year = {2016}, month = {jun}, pages = {174}, month_numeric = {6} }

@article{TCWM062016, author = {Krumpe, Tanja and Walter, Carina and Rosenstiel, Wolfgang and Spüler, Martin}, title = {Asynchronous P300 classification in a reactive brain-computer interface during an outlier detection task}, journal = {Journal of Neural Engineering}, year = {2016}, month = {jun}, volume = {13}, number = {4}, pages = {046015}, month_numeric = {6} }