Initially, the idea appeared as a way to help those with low attention span issues. Given the magnitude of information surrounding us, with around-the-clock Internet access, and the constant shifting between increasingly complex tasks, it is no surprise many of us these days suffer from low attention spans. We thus hypothesized that real-time monitoring of attention level and issuance of feedback may be valuable in situations where attention is paramount, producing improved learning and performance outcomes.

After performing a lot of tests with more than 160 people up to date, we discovered a set of new use cases, scenarios, applications, as well as cognitive processes which our system can measure, like cognitive load or fatigue. We also updated our feedback interventions, now having passive and active options, as well as context-based interventions (if you are falling asleep in the car, a light inside of the car as well as music might be turned on to wake you up along with the suggestion to stop and get some rest).

Since then we moved forward to workplace applications, driving, sports, and other environments where extreme attention, performance, or fatigue are known to occur.

We currently only work with neurotypical adults on this project.

Our primary concern was about solving a problem and not just building a new and fancy device, so for the very first prototype we took an existing hardware from the market. We used a lightweight, commercial EEG headband, similar to the ones used in other studies. Using a band facilitates the reproducibility and generalizability of the study as well as its comparability to previous work. Moreover, the bands have a lower price compared to professional headsets. As one of the core ideas (see question 1) was to also test the feasibility of having a closed-loop system, meaning providing real-time feedback, we needed to provide the users with the feedback about their brain activity. For our work we ruled out the use of visual feedback, as our system is meant to be used in visually charged environments, and we do not want users to get even more distracted by checking their phones or some other devices to view their engagement level. Audio feedback may be a viable option in the case of bone conduction earphones, as such earphones do not require covering the ear. Hence, for our very first study, we decided to use vibrotactile and auditory feedback modalities. Prior research shows that the location of vibrotactile devices on the body determines how effective the vibrotactile stimulation will be. Zeagler as well as Karuei et al. identified the upper chest area below the collar bone as well as the wrist as the optimal body locations. We chose to include the vibration elements in the form factor of a scarf, to be worn on the upper part of the chest. The rationale was to standardize haptic feedback delivery across users and avoid confusion with common, commercial wrist bands they might be used to. We considered several possible configurations of devices to be worn on the upper part of the chest and decided on a scarf as they are made of soft fabrics and thus, the vibrations are less disruptive and more private.

Ultimately, as we tested the device with more than 160 people (all 18+ years old), we moved to the first iteration of the system, incorporating everything within one hardware prototype, keeping two feedback options available, based on the usability tests and comments from the users.

We report and publish only with neurotypical adults (18+ years old) as participants.

The project closely involves users throughout the experiments so as to design and test methods that are socially and ethically acceptable to all parties involved. We are always happy to welcome more collaborators, especially in the domains of psychology, ethics, as well as the use cases we currently target like sports and driving.

An important point to be made about AttentivU is that the system was not developed to be worn 24/7 nor to keep a person continuously attentive or engaged. A continuous high level of attention or engagement is not needed as it can negatively affect learning outcomes. The proposed system might be an option for users who have problems with staying on task when attention is required, e.g., falling asleep at work. As it was suggested by several participants, other use cases are to be considered with this system, such as driving (which we implemented as a driving scenario overnight in a simulator, where users have been tired, not attentive, and even falling asleep. The paper is accepted and will appear in the end of September 2019).

Last but not least, very much needed are discussions about ethical issues with regard to development of such devices, as well as the questions pertaining to the privacy and security of both the data but also of the users of such technology.

No. It has nothing to do with mind reading.

Please see our Research page for the latest updates on AttentivU!

No it is not the same system. We have seen those videos and are very well aware of the potential misuses of such systems. The good news is that 1. we are currently investigating the potential benefits/drawbacks of such systems 2. though our current prototype is a functioning and experimentally validated system, it has already been a 4th iteration of the design. Both its hardware and software characteristics and features have changed dramatically over time. More importantly, the use cases shift and change constantly: with feedback component or without, haptic or auditory feedback, positive reinforcement within context-based interaction or not - we test and try these different versions, as different solutions are suitable for different scenarios and populations 3. Current system is not tested with children, even more so - not in the classrooms 4. We welcome constructive feedback and we are happy to share more information: we have recently had a very big discussion over different social media with general public as well as specialists in ethics and teachers, we have collected a lot of feedback for the use cases regarding tests with minors 5. We welcome external collaborators, especially in the domain of ethics and teaching 6. All our protocols are approved by IRB of the institution. Finally, we fully adhere to the mission statement of our group, and this project is not an exception, please check the statement here.

We welcome external collaborators, especially those who work in ethics domain and teachers! Please contact us at nkosmyna AT media.mit.edu and tell us more!

Cognitive load or cognitive workload is name given to the load on cognitive resources of the mind generated due to cognitive demand of the task. In the context of HCI, cognitive load theory is concerned with the design of information systems that efficiently use users' limited processing capacit during the competition of HCI tasks. In literature, cognitive load has been defined to be of three types, namely intrinsic, extraneous and germane [5].

Mental engagement is a term with multiple definitions but the most common use for it is sustained attention or tonic alertness [1]. The term usually includes energetic arousal on the sleep–wake axis, intrinsic motivation, cognitive performance, and concentration according to Kamzanova et al. [2], Reinerman et al. [3], and Freeman et al. [4]. In our work, we define engagement as the ability to focus on one specific task for a continuous amount of time without being distracted. Moreover, in our studies, we follow a well-documented set of research works in order to increase the reproducibility of our work, as well as comparability to existing papers. For a more extensive discussion, we refer the reader to a review by Oken et al. [1] that includes a detailed discussion of different terms like “alertness”, “arousal”, “sustained attention”, and “vigilance”. We also recommend Freeman et al. [4] and Kamzanova et al. [2] for more in-depth discussions of engagement measures using EEG.

Cognitive fatigue is defined as the unwillingness to continue performance of mental work in alert and motivated participants. Cognitive fatigue can also be defined as a phenomenon characterized by reduction of performance after continuous workload accompanied by subjective experience of exhaustion. It affects different cognitive functions including alertness, working memory, long-term memory recall, situation awareness, judgment, and executive control.

References
[1] Oken, B.S.; Salinsky, M.C.; Elsas, S.M. Vigilance, alertness, or sustained attention: Physiological basis and measurement. Clin. Neurophysiol. 2006, 117, 1885–1901.
[2] Kamzanova, A.T.; Matthews, G.; Kustubayeva, A.M.; Jakupov, S.M. EEG Indices to Time-On-Task Effects and to a Workload Manipulation (Cueing). World Acad. Sci. Eng. Technol. 2011, 80, 19–22.
[3] Reinerman, L.E.; Matthews, G.; Warm, J.S.; Langheim, L.K.; Parsons, K.; Proctor, C.A.; Siraj, T.; Tripp, L.D.; Stutz, R.M. Cerebral blood flow velocity and task engagement as predictors of vigilance performance. Proc. Hum. Factors Ergon. Soc. 2006, 50, 1254–1258.
[4] Freeman, F.G.; Mikulka, P.J.; Prinzel, L.J.; Scerbo, M.W. Evaluation of an adaptive automation system using three EEG indices with a visual tracking task. Biol. Psychol. 1999, 50, 61–76.
[5] Naveen Kumar, Jyoti Kumar, Measurement of Cognitive Load in HCI Systems Using EEG Power Spectrum: An Experimental Study, Procedia Computer Science, Volume 84, 2016, Pages 70-78, ISSN 1877-0509, https://doi.org/10.1016/j.procs.2016.04.068.

AttentivU system in the form-factor of the glasses do not include prescription lenses. Check with your preferred eyewear professional and share these instructions to get prescription lenses for your frames.

Please contact our project lead, Nataliya Kosmyna at nkosmyna @ mit dot edu .

AttentivU platform and Muse platform are very different. AttentivU system relies on eye movements activity (EOG) in addition to brain activity (EEG). We focus on the following mental and cognitive states: fatigue, engagement, cognitive load, which Muse headband does not provide.

Drowsiness detection in the car; internal, external attention, focus, fatigue detection and so on. Please check our Research page to see more examples of the applications.