Analysis of Cognitive Load Using EEG when Interacting with Mobile Devices

Cabañero, Luis; Hervás, Ramón; Gónzalez, Iván; Fontecha, Jesús; Mondéjar, Tania; Bravo, José

doi:10.3390/proceedings2019031070

Cited by 20 publications

(11 citation statements)

References 25 publications

(30 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The literature on cognitive load modelling with EEG and deep learning is recent, not vast and highly scattered [ 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 ]. Most of these models are supervised, which means they require a form of ground truth, usually in task-based categories or task-performance measures.…”

Section: State Of the Art In Cognitive Load Modelingmentioning

confidence: 99%

Modeling Cognitive Load as a Self-Supervised Brain Rate with Electroencephalography and Deep Learning

Longo

2022

Brain Sciences

View full text Add to dashboard Cite

The principal reason for measuring mental workload is to quantify the cognitive cost of performing tasks to predict human performance. Unfortunately, a method for assessing mental workload that has general applicability does not exist yet. This is due to the abundance of intuitions and several operational definitions from various fields that disagree about the sources or workload, its attributes, the mechanisms to aggregate these into a general model and their impact on human performance. This research built upon these issues and presents a novel method for mental workload modelling from EEG data employing deep learning. This method is self-supervised, employing a continuous brain rate, an index of cognitive activation, and does not require human declarative knowledge. The aim is to induce models automatically from data, supporting replicability, generalisability and applicability across fields and contexts. This specific method is a convolutional recurrent neural network trainable with spatially preserving spectral topographic head-maps from EEG data, aimed at fitting a novel brain rate variable. Findings demonstrate the capacity of the convolutional layers to learn meaningful high-level representations from EEG data since within-subject models had, on average, a test Mean Absolute Percentage Error of around 11%. The addition of a Long-Short Term Memory layer for handling sequences of high-level representations was not significant, although it did improve their accuracy. These findings point to the existence of quasi-stable blocks of automatically learnt high-level representations of cognitive activation because they can be induced through convolution and seem not to be dependent on each other over time, intuitively matching the non-stationary nature of brain responses. Additionally, across-subject models, induced with data from an increasing number of participants, thus trained with data containing more variability, obtained a similar accuracy to the within-subject models. This highlights the potential generalisability of the induced high-level representations across people, suggesting the existence of subject-independent cognitive activation patterns. This research contributes to the body of knowledge by providing scholars with a novel computational method for mental workload modelling that aims to be generally applicable and does not rely on ad hoc human crafted models.

show abstract

Section: State Of the Art In Cognitive Load Modelingmentioning

confidence: 99%

Modeling Cognitive Load as a Self-Supervised Brain Rate with Electroencephalography and Deep Learning

Longo

2022

Brain Sciences

View full text Add to dashboard Cite

show abstract

“…In the decision and cognitive science literature, both theta and alpha activity in frontal and parietal regions are commonly linked to measures of cognitive load, i.e., the used amount of working memory recourses (Stipacek et al, 2003;Antonenko et al, 2010;Brouwer et al, 2012), including focused attention and sensory processing (Cabañero et al, 2019). Particularly, spectral theta power has been found to increase with sustained concentration and task difficulty (Gevins and Smith, 2003), while alpha oscillatory activity has been associated with alertness (Kamzanova et al, 2014) and cognitive fatigue (Borghini et al, 2012).…”

Section: Conceptual Frameworkmentioning

confidence: 99%

“Sound” Decisions: The Combined Role of Ambient Noise and Cognitive Regulation on the Neurophysiology of Food Cravings

et al. 2022

View full text Add to dashboard Cite

Our ability to evaluate long-term goals over immediate rewards is manifested in the brain’s decision circuit. Simplistically, it can be divided into a fast, impulsive, reward “system 1” and a slow, deliberate, control “system 2.” In a noisy eating environment, our cognitive resources may get depleted, potentially leading to cognitive overload, emotional arousal, and consequently more rash decisions, such as unhealthy food choices. Here, we investigated the combined impact of cognitive regulation and ambient noise on food cravings through neurophysiological activity. Thirty-seven participants were recruited for an adapted version of the Regulation of Craving (ROC) task. All participants underwent two sessions of the ROC task; once with soft ambient restaurant noise (∼50 dB) and once with loud ambient restaurant noise (∼70 dB), while data from electroencephalography (EEG), electrodermal activity (EDA), and self-reported craving were collected for all palatable food images presented in the task. The results indicated that thinking about future (“later”) consequences vs. immediate (“now”) sensations associated with the food decreased cravings, which were mediated by frontal EEG alpha power. Likewise, “later” trials also increased frontal alpha asymmetry (FAA) —an index for emotional motivation. Furthermore, loud (vs. soft) noise increased alpha, beta, and theta activity, but for theta activity, this was solely occurring during “later” trials. Similarly, EDA signal peak probability was also higher during loud noise. Collectively, our findings suggest that the presence of loud ambient noise in conjunction with prospective thinking can lead to the highest emotional arousal and cognitive load as measured by EDA and EEG, respectively, both of which are important in regulating cravings and decisions. Thus, exploring the combined effects of interoceptive regulation and exteroceptive cues on food-related decision-making could be methodologically advantageous in consumer neuroscience and entail theoretical, commercial, and managerial implications.

show abstract

“…In the context of mobiles, the utilisation of the application may be affected by the capability of users to operate the mobile application and move around. As noted by [5], the mobility of users and the application utilisation of users have to be taken into consideration whilst studying the usability of a mobile. The CL is identified as an attribute of the usability model, PACMAD, without the mentioning of the low-level, related metrics which are representative of each of the attributes.…”

Section: Literature Searchmentioning

confidence: 99%

“…Consequently, this can be achieved when a session-listener is used between the SS (session-creation time) and SE (session-destroy) can be seen in Formula 5: 𝑇𝑇𝐻𝐻𝐻𝐻(𝑇𝑇𝐻𝐻𝐻𝐻. 𝑈𝑈) = ∑ 𝑁𝑁𝑆𝑆𝐻𝐻 𝑆𝑆𝑆𝑆 𝑆𝑆𝑆𝑆 (5) Where NoV denotes the number of visits whilst task (T) is undertaken for each of the users ui within the user set {u1…un}.…”

Section: Clm-dtt Representative Referencesmentioning

confidence: 99%

COG Tool: An Automated Cognitive Measurement of Workload for Mobile Event Logging

Saleh

Enaizan

2022

Int. J. Interact. Mob. Technol.

View full text Add to dashboard Cite

Cognitive workload has been defined as the cognitive processing amount needed by a user in using an application. Cognitive load theory has been gaining increasing levels of importance within usability evaluations for mobile applications and the user interface. The PACMAD model has proven that cognitive workload (CL) can be seen as being a primary usability evaluation attribute in relation to mobile applications. Once the tasks that have been assigned to users within the environment of the mobile application have been performed by the evaluator, he/she subjectively measures the CL. Typically, subjective measurements are taken at a point following completion of a task and so they are static. Since these methods are static in nature, they are inappropriate for assessment of dynamic changes to the cognitive loads through user interaction. Robust measurement of the cognitive load of a user within real-time is not a simple, trivial undertaking. Lots of research studies have tried to assess the CL of users through use of various subjective measurements, though it is often the case that those approaches are unsuitable for use with real-life applications because of the high degree of intrusiveness. Automated measurement is seen as being a field that is relatively unexplored, particularly if used in relation to CL for mobile applications. The COG tool (CL automated tool) is proposed in this paper since it supports the interaction logging of the user for derivation of CL metrics that are based upon both secondary and primary tasks. Several strengths are provided by the automated measurement of CL, and it fills the gaps of a static nature of its non-objective, automated (subjective) counterpart. Within this research, in order for the study aim to be satisfied, objective CL metrics have been used by the researcher, i.e., task time, task success rate, number of touches, number of errors, duration on help, help visit count, total effort percentage and reaction on help. A total of 68 participants were recruited for the experiment, with them using an AAU mobile application through performance of a task that was predefined without there being any interference of a moderator or researcher. The findings showed that the proposed tool could fulfill the expected cognitive load monitoring functions amongst user interaction, collection of cognitive metrics and the logging of CL metrics that could be analyzed further.

show abstract

Analysis of Cognitive Load Using EEG when Interacting with Mobile Devices

Cited by 20 publications

References 25 publications

Modeling Cognitive Load as a Self-Supervised Brain Rate with Electroencephalography and Deep Learning

Modeling Cognitive Load as a Self-Supervised Brain Rate with Electroencephalography and Deep Learning

“Sound” Decisions: The Combined Role of Ambient Noise and Cognitive Regulation on the Neurophysiology of Food Cravings

COG Tool: An Automated Cognitive Measurement of Workload for Mobile Event Logging

Contact Info

Product

Resources

About