Accessible, affordable, and adoptable fNIRS: a Bluetooth® based, mobile-configurable, modular fNIRS system built on the Arduino platform (“NIRDuino”)

Introduction: Hemodynamic functional neuroimaging (HFN) studies have found that drawing statistically significant conclusions about brain-wide associations (across subjects) and ensuring reproducibility (within subjects) requires longer and recurrent sampling of brain activity than current practice allows. To achieve this requirement, affordable, accessible, and adoptable data collection tools are required to collect datasets of the required durations. Functional near-infrared spectroscopy (fNIRS) is a portable hemodynamic neuroimaging modality that is poised to collect large volumes of HFN data due to its portability and relative affordability in comparison to the gold standard: functional magnetic resonance imaging (fMRI) system.

Unfortunately, currently available commercial and open-sourced fNIRS systems struggle from high costs, limited modularity, or limited adoptability due to dependence on wired tethers or personal computers (laptops/desktops). High costs prevent large portions of the global neuroimaging community from collecting and contributing data to enable an inclusive understanding of brain functioning. Limited modularity and wired tethers restrict the regions of interest, imaging resolution, or types of activities that can be studied during experiments. Finally, interfacing only with personal computers (PCs) restricts the variety of researchers that may participate in fNIRS research, since many countries have emerged as mobile-first communities without access to PCs.

To maximize affordability, accessibility, and adoptability, we have built and characterized a wireless (Bluetooth®-based) fNIRS system that can interface directly with any Android mobile device without any additional wires or dongles. The system, named “NIRDuino”, provides users up to eight light emitters and 16 light detectors that can be arranged in any configuration required.

Materials and

Methods: The NIRDuino system hardware has been built on top of the Arduino Nano ESP32 platform that provides a Bluetooth Low Energy (BLE) communication. The light emitters consist of two surface-mounted LEDs with wavelengths of 740nm and 860nm that are illuminated by a variable voltage driver and triggered using NPN-transistor circuitry and a 16-channel multiplexer. The light detectors consist of a large area silicone photodiode with fixed amplification using a precision transimpedance circuit. Analog signals from all 16 detectors are multiplexed, amplified using programmable gain, and digitized using a 24-bit analog-digital-converter. The data from each emitter-detector pair are collected and an additional round of all detector measurements are performed with all emitters turned off to measure dark state. The collected data, and the time elapsed to collect the data, are both wirelessly broadcasted to a custom Android mobile application at a rate of 5 Hz. Each light emitter and detector has been packaged into small PCBs (smaller than a penny) that connect to a dongle using a wired connection. The dongle interfaces with a central controller. The noise and stability of the system were characterized using a cylindrical tissue-mimicking phantom.

Results, Conclusions, and Discussions: The electronics and a protective box for the controller weigh just 110g. These components can easily be interfaced with light blocking fabrics and silicones that can hold the light emitters and detectors to the scalp surface while also anchoring the controller on the head itself. The system has a rechargeable battery life of up to three hours. Each light emitter can output up to 50mW/cm2 of light; each detector provides a gain of up to 140dB. As a result of phantom characterization, it was found that detector can measure as little as 40pW/cm2 of light and has negligible drift ( < 1%) over 30 minutes. These results demonstrate comparable sensing performance to previously published fNIRS systems.

Our system provides an open source, Bluetooth-enabled, and mobile configurable portable fNIRS system for the fNIRS community that they can utilize for their research purposes. This work will ensure researchers globally can participate, contribute, and benefit from neuroimaging research crucial for navigating emerging challenges in mental healthcare.

Acknowledgements (Optional): This research has been supported by the Dorothy J. Wingfield Phillips Chancellor Faculty Fellowship at Vanderbilt University, NIH grants # 1R21MH123873-01 and #R21AG072188, and the Wond’ry (Vanderbilt University’s innovation center).