| Start Date: | 10/14/2020 | Start Time: | 10:00 AM |
| End Date: | 10/14/2020 | End Time: | 12:00 PM |
|
Event Description
BIOMED Master's Thesis Defense
Title:
NeuroHub Fog: Wireless Network Time Synchronization Device for Multimodal Brain Imaging and Hyperscanning Research
Speaker:
Andrew G. Dai, PhD Candidate
School of Biomedical Engineering, Science and Health Systems
Drexel University
Advisor:
Hasan Ayaz, PhD
Associate Professor
School of Biomedical Engineering, Science and Health Systems
Drexel University
Details:
Brain computer interfaces have a variety of applications including but not limited to neuroscience, engineering, computer science, psychology, and rehabilitation. With a wide range of disciplines and advancing technologies, there is a growing interest, especially in using multiple systems concurrently in multimodal/hybrid configurations to extract complimentary aspects of brain activity, and in measuring multiple brains using hyperscanning configurations to investigate brain activities in social interactions. The use of functional neuroimaging in brain computer interface protocols such as functional near infrared spectroscopy (fNIRS) and electroencephalography (EEG) require precise time synchronized transmission of experimental events and acquired data for proper analysis and interpretation. A scalable, portable device that can act as a bridge between multiple monitoring systems with different communication protocols is required for ensuring practicality of these experimental setups. A challenge is presented with the complexity of having multiple brain and body sensors and providing a proper timing of event markers and data acquisition. The original NeuroHub device, developed at Drexel University, offered time synchronization capabilities through four serial ports, a TTL port, and a parallel port. The following generation of NeuroHub was designed as a modular expansion to the original device in order to offer wireless communications to accommodate for modern computing systems with more complex options.
This thesis proposes a solution that will consolidate both previous generations into a single form factor and bridge different brain and body sensors in order to ensure proper synchronizations in time via network communication protocols and physical hardware ports. With the goal of versatility and practicality in mind, the device will ensure improved convenience in the set-up procedures when attempting to coordinate devices made by different companies. This new generation of NeuroHub will also offer a control application to view and configure network and hardware settings. The device will offer reliable timestamping and marker transmission, particularly for brain and body sensors that have different data transmission protocols. In addition to time synchronized transmission of event markers, the new generation of NeuroHub includes logging capabilities that is able to record event marker traffic.
The communication between NeuroHub devices in the network, denoted as the NeuroHub fog, will be maintained by radio link. The transmission of data allows for communication within a wireless local network in order to bridge multiple brain and body sensors. This device reduces the number of cables required in multimodal or hyperscanning setups while maintaining the ability to serve multiple clients. The limitations and burdens of having many cables during these experimental protocols can be improved with network communication, while also preserving the time synchronization capabilities. An additional logging system was implemented to record and monitor the traffic of data for storage and offline analysis. Implementation of a battery also provides enhanced reliability and promotes mobile applications for multimodal and hyperscanning experimental protocols. Mobility of the subjects during experimentation will be increased, in addition to the mobility of the device itself. This implementation also acts as a fail-safe, providing several benefits should the power supply of the computer be compromised. Long lasting, uninterrupted battery operation can allow the devices to continue working in the event of a sudden power outage or in environments where there is no power. In addition, having a battery available can help protect against data loss from potential damage done to the computer by forced shutdowns. Overall, the batteries create the opportunity for each unit to function independently and portably.
This newest generation of NeuroHub consists of a Raspberry Pi 4 Model B with an attached serial port HAT and a 3800 mAh Li battery. Event markers received by the device is broadcasted within the NeuroHub fog utilizing UDP broadcasting capabilities. The device is able to also have bidirectional communication through the attached serial port connected through the UART component. The device is able to be booted in 2 separate modes: DHCP client mode, where the devices within the NeuroHub fog including the NeuroHub itself are assigned dynamic IP addresses by the router and the DHCP server mode, where the NeuroHub is able to assign an IP address to a connected device over the ethernet port. The control application for viewing and configuring settings is implemented over a web server and also includes a manual marker feature.
The verification tests confirmed 100% accuracy in transmitting markers in both single client round trip test configurations and in multiclient configurations. Testing also confirmed the timing of single client round trip times varied from slightly below 10 milliseconds to sub milliseconds based on the configuration. Multiclient configurations confirmed that the time for the NeuroHub to send a marker to a client over the network and receive one back is roughly 4 milliseconds. |
|
Audience:
Undergraduate StudentsGraduate StudentsFacultyStaff |
|
