Source-linked AI summary
The Smartphone Brain Scanner: A Mobile Real-time Neuroimaging System
Arkadiusz Stopczynski, Carsten Stahlhut, Jakob Eg Larsen, Michael Kai Petersen, Lars Kai Hansen
TL;DR
Neuroscience needs mobile devices for quantitative brain measurements in natural settings, and the Smartphone Brain Scanner combines consumer EEG with mobile devices for real-time 3D source imaging. Its evaluation reported an 80–120 ms pipeline delay and motor-area alpha-band averages that compared favorably with laboratory equipment.
Problem
Neuroscience has a recognized need for mobile devices that support quantitative measurements in natural settings.
Source: §1 · p. 1Method
The Smartphone Brain Scanner is a multi-platform mobile EEG framework for real-time 3D source reconstruction from consumer-grade neuroheadsets.
Source: §3 · p. 4, §3.4 · p. 7Results
80–120 ms pipeline delay; motor-area alpha-band averages compared favorably with laboratory equipment, and both showed expected desynchronization during imagined motor actions.
Source: §6 · p. 14Takeaways & Limitations
The authors suggest mobility and simplified application development may open research directions that offset reduced signal quality relative to high-density laboratory equipment.
Source: §6 · p. 14Mobile systems face low-resolution recordings, movement and electrical artifacts, limited computation, and non-real-time operating systems that can affect signal quality and timing.
Source: §1 · p. 2
Abstract
from arXiv · showhide
Combining low cost wireless EEG sensors with smartphones offers novel opportunities for mobile brain imaging in an everyday context. We present a framework for building multi-platform, portable EEG applications with real-time 3D source reconstruction. The system - Smartphone Brain Scanner - combines an off-the-shelf neuroheadset or EEG cap with a smartphone or tablet, and as such represents the first fully mobile system for real-time 3D EEG imaging. We discuss the benefits and challenges of a fully portable system, including technical limitations as well as real-time reconstruction of 3D images of brain activity. We present examples of the brain activity captured in a simple experiment involving imagined finger tapping, showing that the acquired signal in a relevant brain region is similar to that obtained with standard EEG lab equipment. Although the quality of the signal in a mobile solution using a off-the-shelf consumer neuroheadset is lower compared to that obtained using high density standard EEG equipment, we propose that mobile application development may offset the disadvantages and provide completely new opportunities for neuroimaging in natural settings.
1 Introduction
The introduction motivates mobile neuroscience as a way to measure brain activity in natural settings and presents the Smartphone Brain Scanner as a portable platform for acquisition, processing, and real-time imaging. It weighs that opportunity against technical and signal-quality challenges in consumer-grade mobile systems.
1 Introduction
- Consumer neuroheadsets have made mobile brain monitoring feasible, while smartphones and tablets can sense, process, transmit, and present information.The introduction frames the Smartphone Brain Scanner as a platform for stimulus delivery, data acquisition, processing, and real-time brain imaging.Source: §1 · p. 1
- Mobile EEG systems offer portable, lower-cost, and more comfortable recording that may support longer studies and multi-subject research in natural settings.Built-in wireless transmission and improved comfort can extend experiments over several hours; the authors suggest low cost may open social neuroscience studies involving interacting subjects.Source: §1 · p. 2
- Consumer EEG systems face computational, timing, synchronization, and signal-quality challenges, including low-resolution recordings and movement-related artifacts.The authors describe mobile devices’ limited computational power and non-real-time operating systems, as well as potentially suboptimal electrode placement compared with standard EEG.Source: §1 · p. 2
- The authors frame mobile EEG either as a stand-alone low-fi solution or as a complement to standard laboratory neuroimaging in natural conditions.They argue that the drawbacks may be offset by studying larger groups over longer periods in more natural settings.Source: §1 · p. 2
- A C++ application framework, approached as a smartphone-sensing challenge, may attract new neuroscience developers and enable new application contributions.The authors also point to emerging interest from HCI researchers and general tinkerers.Source: §1 · p. 2
2 Related Work
The related work places mobile brain imaging between neuroimaging and smartphone-based cognitive monitoring. It reviews existing EEG software and consumer sensors, noting both mobile-development gaps and opportunities for extended monitoring.
2 Related Work
- The Smartphone Brain Scanner is positioned between a downsized neuroimaging device and a smartphone sensor system for cognitive monitoring in natural conditions.The authors review work in both domains.Source: §2 · p. 2
2.1 Neuroimaging
- Existing EEG toolboxes and BCI frameworks support analysis or real-time applications, but the reviewed platforms are not readily suited to mobile development.EEGLAB and FieldTrip are MATLAB toolboxes; OpenViBE and BCI2000 target desktop operating systems, while the reviewed SDKs are mainly aimed at immediate-feedback applications.Source: §2.1 · pp. 2–3
2.2 Cognitive Monitoring Systems
- Consumer neuroheadsets enable mobile EEG monitoring, and the framework can record for hours, but solution-based electrodes may dry out before continuous all-day use.A Galaxy Note battery test recorded and stored Emotiv EPOC data for 11 uninterrupted hours with wireless radios and screen off; other physiological and wearable sensors also contribute to cognitive monitoring.Source: §2.2 · pp. 3–4
3 Methods: The Smartphone Brain Scanner
SBS2 is a multi-platform framework for research and end-user EEG applications, organized into acquisition, processing, and application layers. Its modular mobile architecture supports real-time EEG processing, including 3D source reconstruction.
3 Methods: The Smartphone Brain Scanner
- SBS2 is a software platform for multi-platform EEG applications, divided into data acquisition, data processing, and application layers.The first two layers form the core shared by different applications.Source: §3 · p. 4
3.1 Key Features
- The Qt-based framework targets desktop systems and mobile devices, with support for Android, BB10, and partially iOS.Modules can be added or modified, and applications support real-time use as well as recording for later offline analysis.Source: §3.1 · pp. 4–6, Figure 1 · p. 5
3.2 Data Acquisition
- EEG-specific mounters, readers, and packet creators abstract hardware details, while the current Android EPOC setup requires USB host support, HIDRAW, and root privileges.The EPOC data stream is encrypted and the acquisition layer includes a decryption module.Source: §3.2 · p. 6
3.3 Data Processing
- The hardware-agnostic processing layer dispatches EEG packets for recording, filtering, reconstruction, and feature extraction, making raw values and features available to applications.Implemented operations also include FFT, spatial filtering, classification, and network streaming.Source: §3.3 · pp. 6–7, Figure 3 · p. 7
3.4 3D Imaging
- Source reconstruction estimates the brain sources most likely to produce scalp EEG, addressing an ill-posed inverse problem by imposing prior information.SBS2 implements Bayesian Minimum-norm and LORETA formulations, with hyperparameters adapted in real time to different noise environments.Source: §3.4 · pp. 7–8
- The source model relates scalp signals to cortical sources through a forward matrix; it assumes normal, uncorrelated, time-independent noise and uses a graph-Laplacian spatial prior with smoothness 0.2.The paper describes the Laplacian prior as ensuring spatial coherence between sources.Source: §3.4 · p. 8
4 Methods: Experimental Designs
The experiments evaluate EEG timing and data quality with the Emotiv headset, then demonstrate a mobile imagined-finger-tapping task. They assess sampling rate, packet quality, system delay, and task design.
4.1 Timing and Data Quality
- The timing and data-quality experiments assess synchronization needs for paradigms such as P300 and steady-state visual evoked potentials, while also considering extended recordings where precise synchronization matters less.The section frames these experiments as validation of the SBS2 framework, hardware, and mobile approach.Source: §4.1 · p. 8, §4 · p. 8
4.1.1 Emotiv EEG sampling
- The Emotiv sampling-rate validation tested three randomly selected devices with ten 10-minute measurements per device against the nominal 128Hz rate.The headset internally samples at 2048Hz and downsamples to 128Hz.Source: §4.1.1 · p. 9
4.1.2 Data Quality
- The data-quality assessment tracks dropped packets with a modulo 129 counter and reports zero packet loss during a full rundown recording with good transmitter-receiver visibility.The headset provides channel-quality information; its manufacturer’s green indicator corresponds to a channel-quality value greater than 407 and is reported to indicate sufficiently low impedance.Source: §4.1.2 · p. 9
4.1.3 Timing
- System delay was measured from a 10Hz signal reaching the EEG hardware until a detected peak changed the device screen from black to white.A photocell detected the screen change, enabling oscilloscope measurement of delay and jitter.Source: §4.1.3 · p. 9
4.2 Imagined Finger Tapping
- The imagined finger-tapping experiment used randomized left and right cues, a central fixation point, and 200 trials from one subject.The display showed Relax, Left, and Right instructions; the subject was asked to focus on the screen center to minimize eye movements.Source: §4.2 · p. 10
5 Results and Discussion
The online source reconstruction produced similar left-dominant alpha suppression for imagined right tapping with Emotiv and Biosemi recordings. Sampling and timing measurements also indicate that the headset’s rate is consistent and its system delay is stable enough to correct.
5.1.1 Emotiv EEG sampling
- The Emotiv EPOC sampling rate was close to 127.88Hz and fairly consistent; analyses may assume 128Hz or 127.88Hz, or measure each device’s rate.Source: §5.1.1 · p. 10
5.1.2 Timing
- 80–125ms of system delay was fairly stable, with 16–26ms jitter, and therefore could be corrected for.Figure 6 reports device-specific timings for the Galaxy Note, Nexus 7, and MacBook Pro.Source: §5.1.2 · p. 11, Figure 6 · p. 11
5.1.3 3D Source Reconstruction On-Device Performance
- On-device source reconstruction and hyper-parameter calculation were tested on 1s of signal, with reported timings of 2ms/2s for MacBook Pro, 8ms/1s for Nexus 7, 8ms/11s for Galaxy Note, and 14ms/13s for Acer Iconia.Source: §5.1.3 · p. 11
5.1.2 Timing
- Packet timing measurements indicate that the Emotiv system compensates every 32 samples to keep the average interval at the expected level.The measurements changed the screen color every four packets and used a photocell to assess the intervals.Source: §5.1.2 · p. 11, Figure 7 · p. 12
5.2 Imagined Finger Tapping – Online Source Reconstruction
- The left-dominant alpha suppression for imagined right finger tapping looked similar with Emotiv EPOC and Biosemi, with source estimates and model parameters calculated online.The responses were averaged across 87 right-cued trials for source localization and 79 right-cued trials for power response.Source: §5.2 · p. 12
6 Conclusions
The Smartphone Brain Scanner combines off-the-shelf EEG hardware with Android phones and tablets for fully mobile 3D imaging. Evaluation showed stable, low-delay imaging and motor-area signals that compared favorably with laboratory equipment, while the authors note the expected lower signal quality of mobile consumer headsets.
6 Conclusions
- The open-source Smartphone Brain Scanner is presented as the first fully mobile 3D EEG imaging system.It supports real-time EEG data acquisition.Source: §6 · p. 12
- Standard Android smartphones and tablets perform real-time source imaging at frame rates above 40 fps, using a solver with online Bayesian optimization of noise-level and regularization hyperparameters.Source: §6 · p. 14
- 80–120 ms delay and aggregate alpha-band signals that compare favorably with laboratory equipment support a stable imaging pipeline, with expected desynchronization during imagined motor actions.Figure 8 shows reconstructed source power in left and right precentral regions across right-cued imagined tapping conditions, with online α and β estimation using a Minimum Norm Solution.Source: Figure 8 · p. 13, §6 · p. 14
- The authors propose that mobility and simplified application development may offset the lower signal quality of consumer low-density neuroheadsets relative to controlled, high-density laboratory equipment.Source: §6 · p. 14