Event Abstract Back to Event EEG-based neural correlates of ACT-R model for multitasking Nayoung Kim1, Erica McCune1, MyungHwan Yun2 and Chang S. Nam1* 1 North Carolina State University, United States 2 Seoul National University, Republic of Korea Adaptive Control of Thought-Rational (ACT-R) is a high-level computational simulation of human cognitive processing, and one of cognition theories that seek to predict human performance in cognitive tasks such as multitasking (Salvucci et al., 2014). Comparing simulated human cognitive data from the ACT-R to real human performance data has shown to provide important insights into the cognitive mechanisms behind human cognition such as how brain networks are affected by changes in internal cognitive strategies, external interface properties, and task demands. ACT-R data and fMRI activity have been effectively compared to show a clear relation between the two (van Vugt, 2012). fMRI research has helped us to understand the roles of various brain regions and assigned the cognitive functionality of those regions to ACT-R modules and buffers (Cassenti et al., 2011). Despite this strong correlation between fMRI data and ACT-R, it is still unknown whether ACT-R also holds a strong correlation with EEG data. Although fMRI provides valuable spatial data, it measures changes in blood flow which are indirect markers of brain electrical activity. On the other hand, EEG which is the brain's direct electrical activity has a superior temporal resolution (i.e., on the order of milliseconds) compared to fMRI and thus could provide insights into the differences in the time course of ACT-R module activation as well as the modules’ interaction (van Vugt, 2014). Recently, researchers have used various methods to investigate the neural correlates of ACT-R in electrophysiological data and demonstrated how different ACT-R modules can be associated with observable EEG data (van Vugt, 2012; 2014). Despite these initial attempts to correlate ACT-R and EEG, this area still requires more investigation. The present study aimed to extend these initial findings by further exploring EEG correlates of ACT-R modules and discuss the broader implications of this approach for both neuroergonomics and cognitive modeling with ACT-R. Specifically, this study developed models of multitasking behavior in a realistically complex workspace, incorporating a wide range of cognitive motor processes that are affected by workload transitions. Previous studies have shown that changes in task difficulty leads to changes in cognitive demand and task performance (e.g., Bowers et al., 2014). Multitasking environments can often give rise to such transitions in workload, as the operator is actively monitoring and engaging in several different tasks at once. To better understand this process, a computational model of multitasking was developed and its performance was juxtaposed against human operators performing the same task. Twelve trained participants (7 male and 5 female with mean age of 23.17) were recruited from a local university. ACT-R data were recorded within a JavaScript implementation of a modified multi-attribute task battery (mMATB-JS, Cline et al., 2014). This web-based version of the MAT-B contains up to four simultaneous tasks that are the same as the Air Force Multi-Attribute Task Battery (AF-MATB, Miller, 2010). The AF-MATB is a computer-based task designed to evaluate operator multitasking performance where participants perform a tracking task while concurrently monitoring warning lights and dials, responding to computer-generated auditory requests to adjust radio frequencies, and managing simulated fuel flow rates using various key presses. Performance in each subtask was individually scored. During this task, EEG signals were recorded using an EEG cap (Electro-Cap International, Inc.) embedded with 62 active electrodes, based on the modified 10–20 system. Recordings were referenced to the left ear lobe and grounded to AFz. EEG signals were amplified with a g.USBamp amplifier (g.tec Medical Engineering) and sampling rate was 256 Hz. Independent component analysis was used to decompose the EEG signal into independent components (ICs). Signal acquisition and processing were all conducted using the BCI2000 system (Schalk et al., 2004) and EEGLAB (Delorme et al., 2011). In the mMATB-JS, all task parameters were matched to AF-MATB setting. Task demands were manipulated by increasing the event rate in each of the subtasks. Three conditions were used with a fixed order: Low-High- Low. We examined the utility of multitasking behavior when comparing human-to-model data. To compare performance of human and the ACT-R model, Independent Components (ICs) derived from EEG-data were linked to ACT-R buffer activation, using dipole fitting and brain effective connectivity analysis. Prezenski et al (2016) demonstrated that EEG data could be used to validate ACT-R models. If the ICs match specific ACR-R modules, the timing of peaks of buffer activity should match IC-peaks. However, in this study, we observed both IC peaks timing and brain connectivity between ICs. In the ACT-R model, these modules are designed to occur in specific areas of the brain: Imaginal in the parietal cortex, Declarative in the DLPFC, Goal in the ACC and Visual in the fusiform gyrus. Once the electrophysiological correlates of ACT-R are determined by goodness of fit, interaction between multiple modules within the ACT-R model can be monitored by analyzing patterns of synchronization in brain network by using SIFT toolbox (Mullen, 2010). For the first Low 1 condition, we found strong information flows between the visual cortex (VC) and dorsal anterior cingulate cortex (dACC), which matched a link to the goal module and visual module within ACT-R model. Following the first transition of workload (low- high), the causality flow was observed between the parietal cortex (PC) and dorsal anterior cingulate cortex (dACC), which indicates a connection between imaginal modules and goal modules. Lastly, after the second workload transition (high-low), we found the causal flow from prefrontal cortex to dACC which indicated connectivity between declaratives modules and goal modules. By monitoring the interactions between modules, findings of this study will improve understanding of how different brain regions interact within the ACT-R model and could be used to enhance existing cognitive models. Brain network analysis would allow us to interpret causal relationships and efficiency across cognitive model components and specific brain regions. We plan to develop this multitasking model to account for applied workload transition effects and do further real-time analysis to integrate brain network dynamics and model development. Continued interdisciplinary research of cognitive modeling and neuroscience will lead to better understanding of cognitive processes of the human brain at work. Acknowledgements This work was supported in part by the National Science Foundation (NSF) under Grants IIS-1421948 and BCS-1551688. This research was also supported in part by Brain Pool program funded by the Ministry of Science and ICT (MSICT) through the National Research Foundation of Korea (2018H1D3A2001409). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF or MSICT. References Bowers, M. A., Christensen, J. C., & Eggemeier, F. T. (2014). The effects of workload transitions in a multitasking environment. Proceedings of the Human Factors and Ergonomics Society, 2014–Janua, 220–224. Cassenti, D. N., Kerick, S. E., & Mcdowell, K. (2011). Observing and modeling cognitive events through event-related potentials and ACT-R. Cogn. Syst. Res., 12(1), 56–65. doi:10.1016/j.cogsys.2010.01.002 Cline, J., Arendt, D. L., Geiselman, E. E., & Blaha, L. M. (2014). Web-based implementation of the modified multi-attribute task battery. In 4th Annual Midwestern Cognitive Science Conference, Dayton. Cox-Fuenzalida, L.-E. (2007). Effect of Workload History on Task Performance. Human Factors: The Journal of the Human Factors and Ergonomics Society, 49(2), 277–291. https://doi.org/10.1518/001872007X312496 Delorme, A., Mullen, T., Kothe, C., Akalin Acar, Z., Bigdely-Shamlo, N., Vankov, A., & Makeig, S. (2011). EEGLAB, SIFT, NFT, BCILAB, and ERICA: New tools for advanced EEG processing. Computational Intelligence and Neuroscience, 2011. https://doi.org/10.1155/2011/130714 Miller, W. D. (2010). The U.S. Air Force-Developed Adaptation of the Multi-Attribute Task Battery for the Assessment of Human Operator Workload and Strategic Behavior. Retrieved from http://www.dtic.mil/docs/citations/ADA537547 Morgan, J. F., & Hancock, P. A. (2011). The effect of prior task loading on mental workload: An example of hysteresis in driving. Human Factors, 53(1), 75–86. Prezenski, S., & Russwinkel, N. (2016). A proposed method of matching ACT-R and EEG-Data. In In Proceedings of the 14th International Conference on Cognitive Modeling, eds D. Reitter, and FE Ritter (University Park, PA: Penn State) (pp. 249-251). Salvucci, D. D., Kushleyeva, Y., & Lee, F. J. (2004). Toward an ACT-R General Executive for Human Multitasking. In ICCM(pp. 267-272). Schalk, G., McFarland, D. J., Hinterberger, T., Birbaumer, N., & Wolpaw, J. R. (2004). BCI2000: A general-purpose brain-computer interface (BCI) system. IEEE Transactions on Biomedical Engineering, 51(6), 1034–1043. Schwiedrzik, C. M., Ruff, C. C., Lazar, A., Leitner, F. C., Singer, W., & Melloni, L. (2014). Untangling perceptual memory: Hysteresis and adaptation map into separate cortical networks. Cerebral Cortex (New York, N.Y. : 1991), 24(5), 1152-1164. doi:10.1093/cercor/bhs396 van Vugt, M. K. (2012). Relating ACT-R buffer activation to EEG activity during an attentional blink task. van Vugt, M. K. v. (2014). Cognitive architectures as a tool for investigating the role of oscillatory power and coherence in cognition. NeuroImage, 85, 685-693. doi:10.1016/j.neuroimage.2013.09.076 Keywords: EEG, ACT-R, brain connectivity analysis, Cognitive Modeling, multitasking Conference: 2nd International Neuroergonomics Conference, Philadelphia, PA, United States, 27 Jun - 29 Jun, 2018. Presentation Type: Oral Presentation Topic: Neuroergonomics Citation: Kim N, McCune E, Yun M and Nam CS (2019). EEG-based neural correlates of ACT-R model for multitasking. Conference Abstract: 2nd International Neuroergonomics Conference. doi: 10.3389/conf.fnhum.2018.227.00060 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 02 Apr 2018; Published Online: 27 Sep 2019. * Correspondence: Prof. Chang S Nam, North Carolina State University, Raleigh, United States, csnam@ncsu.edu Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Nayoung Kim Erica McCune MyungHwan Yun Chang S Nam Google Nayoung Kim Erica McCune MyungHwan Yun Chang S Nam Google Scholar Nayoung Kim Erica McCune MyungHwan Yun Chang S Nam PubMed Nayoung Kim Erica McCune MyungHwan Yun Chang S Nam Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.