The recent emergence of consumer-grade wearable electroencephalographic (EEG) systems has opened potential new avenues for research and investigation into the understanding of the human brain through postnatal development [1] and neurological disease [2, 3]. Unlike traditional EEG devices that are largely confined to hospitals and research settings, brain wearables are portable, battery-operated, and require just a few minutes for setup without the need for a modality-specific technician [4]. Because of these practical advantages, coupled with significantly lower cost (~ $200–900 USD/device [5]), there is hope that these devices may facilitate detection and monitoring of neurological pathologies outside of traditional centralized research and healthcare settings [6, 7]. Despite their accessibility and ease of use, the integration of such wearable devices into large-scale longitudinal research outside of clinical settings remains largely unexplored due to the absence of a suitable framework for standardized data collection and analysis.Previous work using the four-electrode “Muse 2” EEG headbands has shown that these consumer-grade devices can capture key EEG analysis metrics, such as event-related potential (ERP) and resting state power spectral data that are comparable to clinical-grade EEG systems [8, 9]. Moreover, using three-minute recordings from the same devices, others have shown the ability to predict stroke severity by comparing the power spectral properties of ischemic stroke patients and healthy controls [10]. Other hardware like the Emotiv EPOC + Saline Flex have also been validated for the purposes of ERP collection using an approach involving simultaneous recordings with a research-grade system [11]. While these important pioneering studies demonstrated the potential utility of consumer devices for cognitive monitoring and biomedical applications, they were largely confined to the research setting and were not designed to leverage their unique portability and operator-free benefits.Other recent studies have used remote data collection protocols to demonstrate the practicability of gathering frequent at-home data from participants. For example, one group deployed the Muse 2 device to collect daily remote data over 14 days to study mindfulness and found correlations with self-report mind wandering metrics [12]. More recently, Sidelinger et al. validated remote and longitudinal spectral resting state EEG data from the Muse 2 by collecting data with a proprietary software. When they compared the wearable data to in-lab medical grade EEG recordings, they found significant correlation between self-reported trait anxiety and day-to-day variability of an individual’s alpha frequency [13]. The most popular application of at-home EEG are those designed to mimic sleep studies with polysomnography. For example, one group used connected an ear-EEG sensor to a portable amplifier, but required a technician to visit the users home [14]. Similarly, another study developed an at-home sleep-staging and apnea-detecting device, but also equipped the device in the presence of a technician [15]. These innovative works overcame the barriers associated with remote EEG collection, but there remains a clear need for an open-source system to allow participants to equip the devices independently and use them to carry out advanced neurocognitive assessments.There have been various attempts to develop platforms for remote data collection, each with a different design focus. An early system known as “NeuroMonitor” allowed a small circuit board and wired electrodes to transmit data with Bluetooth [16]. Li et al. developed a similar portable EEG system that suppressed external noise sources and collected robust signal but still involved a circuit board being strapped to the user’s arm, prohibiting independent use by a patient population [17]. An open-source project known as “cEEGrid” developed a small wearable device that leveraged OpenBCI EEG signal acquisition platform, which is not adaptable to other commercially available hardware [18]. Milne-Ives et al. reviewed the subset of the available monitoring technologies for people with epilepsy, and found that few systems were both adequately reported and sufficiently capable [19]. One popular mobile application known as “Mind Monitor” has been used in many publications but this software is limited to the Muse devices and does not allow stimuli to be synchronized [20, 21]. Other commercial software like Interaxon’s “Muse Direct” suffers from a lack of synchronized event markers and others like “Emotiv Pro” have expensive subscription models [22, 23].Notwithstanding these important milestones, there is still a clear need for an open-source companion platform to allow for implementation of cognitive tasks for the measurement of event-related potentials (ERPs) in the remote setting.6 Because ERPs have been shown to be informative for a variety of pathologies, the ability to monitor them remotely has important implications for the scalability of neurocognitive disease research. The current lack of a suitable platform for these purposes prohibits the deployment of EEG wearables toward novel avenues for the real-time and longitudinal study of disease detection and evolution. To address this critical gap, we have developed “HEROIC” (Home EEG Recording frOm Interfacing Computer), an open-source research platform capable of leveraging consumer-grade EEG wearables to longitudinally deploy a battery of cognitive process tasks to measure brain activity (including ERPs [9, 24,25,26]) in both research and remote settings. As a proof-of-concept, we deployed our system to record four at-home sessions from 14 healthy participants resulting in a dataset containing approximately 60 independent EEG recording sessions. We use these data to highlight that HEROIC is easy-to-use and reliable for collecting high quality data and support its ability to be deployed remotely and longitudinally to measure brain activity including complex markers like ERPs. We make our unique software and sample dataset publicly available to provide a democratized platform for the research community to use for the investigation of brain health and disease.In addition to a brief introduction of the HEROIC platform, we describe its implementation and how it can be operated to collect EEG data. We also describe a proof-of-concept pilot study as an example of how a researcher could design a protocol that uses HEROIC. Finally, we will present the results and analysis of this feasibility study demonstrating that HEROIC is capable of remotely and independently measuring precise quantitative brain activity.
