A.02._4.18.19_Amendment_0001_RFI_response_to_question_for_fNIRS.docx

RFI response for fNIRS/EEG system

Vendor comment: We have some questions related to this RFI. Overall, the language in this RFI is very much targeted towards the Artinis Dual Brite system (Artinis is our competitor, Dual Brite is their specific system) and the TMSi Mobita sytem, and the functional reasons for the technical requirements are not clear. We do not have an issue with many of the requirements, though it would be helpful to understand the underlying research needs that yield certain less-common and research-specific requirements that can be met with various technical capabilities, which is where our questions are coming from.

Response: The general requirements are based on Artinis equipment as a template, but we are not necessarily looking for a certain band name. Wherever a requirement is listed that would be proprietary to a certain provider, another named product that will meet or exceed the performance for that requirement is acceptable.

Vendor Comment 2.a: Must be capable of near whole head coverage with variable configurations of optodes. How is “near whole head” defined? Would it be preferred to have whole head coverage or would that be too much? Please clarify the exact requirement and why it is required. Are there any particular brain regions of interest that will be measured? Are all measurable areas of interest?

Response: Near whole head for this requirement is defined as covering as much of the cortex as possible. During market analysis for this requirement several systems are capable of covering the majority of cortex space. Complete coverage from orbital-frontal to cerebellum seems unlikely. Listing the requirement to “near whole head” is to convey that we wish to be able to measure as much of the cortical space as possible and be able to adjust optodes as necessary based on the specific research question. Regions of interest will change as a result of the research being conducted, but most projects will be looking for network connectivity and strength (i.e., default mode network, attention networks, fronto-parital connectivity, or semantic processing centers) and not a specific ROI so in general, the vendor is correct that all measurable areas are of interest.

Vendor Comment 2.b: Must have high volume of channels equaling 54 or higher, and consisting of 20 transmitting optodes and 16 receiving optodes. How are channels counted? In fNIRS, channels may refer to any of the following, for example: only “nearest neighbor connections” (topography), each wave length of connection, or tomographic data (tomography). Does this include short channel data (used for removing superficial skin blood flow)? If the channel requirement is met, why is there also a transmitting optode (and receiving optode requirement? That is, this minimum channel requirement can be met with a various numbers of transmitting optodes and receiving optodes. Why are20 sources and 16 detectors functionally required? Please clarify.

Response: The exact number of electrodes is not necessarily important. What is important is the coverage of cortical space and the quality of data. If there is another optode number/configuration that will meet or exceed the cortical coverage (not exceeding the optimal 3cm per channel) and quality of data, then we would consider this requirement met. Short channel data is not specifically accounted for in the minimum specification list, however, if this is a capability that exceeds the minimum requirements and can be included in the purposed package, then we would be interested in having that capability.

Vendor Comment 2. d: Sampling rate must be 50 Hz or higher for all channels employed. Please clarify the purpose of this requirement: that is what is the physiology being measured that would require 50 Hz sampling rate with fNIRS? Typical fNIRS signals analyzed are below 10Hz.

Response: Sampling rate is listed to emphasize data acquisition quality. If another system meets or exceeds all other requirements and provides high quality data acquisition at a lower sampling rate, then that system should be considered.

Vendor Comment 2. e: Must be able to integrate 9 axis orientation sensor within main unit. Please clarify why the orientation sensor must be within the main unit? The main unit for the Brite system is actually on the head, is this why this is preferred? What is the functional reason for the orientation sensor being inside the main unit? Would it not be preferred to have orientation sensor at the position that the data are collected, even if it is outside the main unit?

Response: The capability for 9 axis orientation is what is important. Where that capability resides (in main unit or not) is not as important.

Vendor Comment 2. J: Package needs to include laptop with all software needed for data acquisition and data analysis of fNIRS EEG data with appropriate licenses such as OxySoft for NIRS and EMSE for EEG. These packages are proprietary and only distributed by Cortech in the United States. Are there functional aspects of these packages that are desired that can be perhaps offered by other manufacturers or are you only interested in considering systems from Cortech/Artinis? OxySoft for NIRS: fNIRS recording and analysis. What features are required/desired? EMSE for EEG: EEG recording and analysis. What features are required/desired?

Response: The exact name of the software is not important. Army network system security requirements prevents the use of most free-ware applications that are used to analyze research data (for example FSL for MRI data analysis cannot be on an Army system). Therefore we require all the software needed to collect and analyze fNIRS and EEG data. Other software packages such as Turbo Satori and BrainVision Analyzer would meet this requirement. We have the laptop as a requirement so that we would have a laptop with a configuration (i.e. memory and processing speed) that meets the requirements for the software and be dedicated to the system. Multiple operators and off site data acquisition requires a dedicated computer for the system to avoid the use of a normal workstation computer doing double duty.

Vendor comment 3.f: Must be capable of variable sampling frequencies as high as 2000 Hz and as low as 500 Hz. What is the reason for the functional requirement of sampling up to 2000 Hz? What are the specific signals and function being acquired and investigated that would require anything more than a 500 Hz sampling frequency? For reference, typical EEG frequencies acquired in open environments are below 50Hz, anything above this is very difficult, if not impossible to analyze. Please clarify this requirement.

Response: Same as response to Vendor Comment 2. d above.

Vendor comment 4. b: Package software must offer Lab Streaming Layer support for the transmission of physiologic data and TTL triggers. Also must offer MatLab FieldTrip plug in for real-time streaming of data. Would other MATLAB or proprietary plug-ins for real-time streaming of data be acceptable?

Response: Yes

Other Questions:

Is it required to be able to use all fNIRS optodes and EEG electrodes simultaneously in recordings?

Not necessarily. The research question will dictate if we need to focus on EEG or fNIRS. The desire is to have the most coverage of both for simultaneous data acquisition. We do understand that within systems, there is always a tradeoff. In this case, the maximum number of optodes will likely restrict the number of electrodes that will fit with the reverse being true as well. We want the maximum in flexibility and capability as it impossible to anticipate what the Army will want us to research in the future and we need a versatile research tool.

Is it required to use EEG in a ‘standard’“10/20” EEG layout during simultaneous fNIRS/EEG recordings?

Within reason. The standard 10/20 system is widely accepted in the literature and deviations from that will likely be questioned by reviewers. We did anticipate that electrode placement may not exactly fit into the 10-20 layout so we included in the package requirements that the package must contain the means of acquiring x, y, z coordinates of optodes/electrodes for reporting and reproducibility.

Are short-distance detectors required for being able to measure superficial blood-flow in the scalp, allowing for isolating and identifying brain activity?

There is no specific requirement listed. However, if this is a capability that a package can provide, that capability would be considered when evaluating a package.

Is there a requirement to be able to upgrade to simultaneous whole-head coverage with fNIRS optodes spaced over entire head during a single recording?

We require the capability to cover as much of the cortical surface (i.e., near whole head) as we can with fNIRS and EEG for simultaneous data acquisition in a single recording. fNIRS and EEG signals may be collected using different processes, but they must be collected simultaneously with the ability to trigger or mark both data streams for syncing data and marking task related time points.

Are there any requirements related to measuring in open environments?

Yes. I am unsure what is meant by open environment, but I assume it refers to non-laboratory space. If the assumption is true, we do require measuring in open environments which resulted in the wireless or Bluetooth recording capability requirement so that we are able to collect data in various locations such as inside vehicles, simulators, or even shooting ranges.

Are there any requirements related to EEG or fNIRS data quality in electrically noisy environments?

There is no requirement for the package to be MRI compatible. However, we do plan to use the system in various environments such as inside vehicles (ground or air), simulators, and possibly shooting ranges.

Are there any requirements related to EEG or fNIRS data quality during subject movement?

We understand excessive movement will degrade data quality. However, our participants will likely have some movement based on the task. We require wireless or Bluetooth data acquisition for this purpose. For example, a pilot conducting a task in a simulator will move their head around, but do not anticipate needing to collect data on a subject that is running and jumping.

Would active electrode technology with pre amplification be preferred?

I am not familiar with pre amplification. If there is a benefit, then we would gladly take that into consideration when evaluating a package.