Attachment 1 21-001953 - Purchase Description.pdf

Version 11-30-2016-1 FAR Part 11—Describing Agency Needs

PURCHASE DESCRIPTION

1. Statement of Need and Purpose:

There is currently a need in the field of non-invasive neuroimaging to develop a technique that can map the brain at both high temporal resolution and high spatial resolution. While fMRI has spatial resolution on the order of 1mm, the temporal resolution is generally in the range of 1-2 seconds and is further impacted by the slow hemodynamic response to stimuli. MEG, in contrast, has extremely high temporal resolution, but spatial resolution in the 5-10mm range. Our overarching aim is to develop a system that has spatial resolution on the order of 1mm, with the temporal resolution of MEG. Under a previous award, we built a prototype optically pumped magnetometer (OPM) magnetoencephalography (MEG) system consisting of 16 detection channels and an additional 3 reference channels.

With the acquisition described here, we plan to expand this prototype 16 channel system to a 56 channel system, capable of biomagnetic imaging at a spatial resolution comparable to that of invasive intracranial electrocorticography (ECoG), over a localized area of the cortex.

2. Background Information and Objective:

The mission of the NIMH MEG Core Facility is to maintain a state-of-the-art MEG platform to obtain in vivo measurements of brain function. The MEG core facility currently possesses a state-of-the-art CTF 275 channel MEG whole-head system and a prototype 16-channel OPM system. Current commercially available MEG systems are able to detect extremely weak electrical fields (on the order of femto-tesla) using superconducting quantum interference devices (SQUIDs). These sensors must be maintained at an ultra-low temperature by immersion in liquid helium, and thus cannot be placed in immediate proximity to the research participant’s head.

In contrast, OPM sensors do not require cooling, and can be placed such that the sensitive region of the sensor is only a few millimeters from the head. Because the field of a dipolar current source declines as 1/r2, where r is the distance from the source, the distance between the cortical sources and the sensors is of paramount importance.

Additionally, the source to sensor distance determines how well the higher spatial frequencies of the brain’s magnetic fields are sampled by the sensor array. Merely decreasing the spacing between sensors is not sufficient to discriminate among multiple nearby cortical sources, unless the distance from the sensors to the cortex is reduced.

This explains why the SQUID-based MEG instruments, while well suited for whole head measurements, cannot achieve source resolution obtained using invasive electrodes.

To summarize, the performance of an array of magnetometers to resolve independent closely-spaced cortical sources depends upon three factors: 1) distance between the magnetometer sensor and the cortical sources, 2) inter-sensor spacing, and

3) signal-to-noise ratio (where noise is defined as the combined brain noise, environmental noise, and sensor noise). We, therefore, aim to develop and evaluate an array of OPM magnetometers in order to achieve source resolution on par with that of invasive cortical grids.

Experimental OPM sensors have demonstrated sensitivities of 5 fT/√Hz, on par with that of SQUID sensors. Simulation studies comparing simple OPM arrays to SQUID arrays have shown that reducing the distance to the head can improve signal-to-noise ratio (SNR) by a factor of 5-7, and localization accuracy by more than a factor of 2.

Published and unpublished simulations have demonstrated that 1 mm resolution is readily achievable with an on-scalp system. Advances in source reconstruction may improve upon this, as it has recently been demonstrated that isolation of MEG signals to either the pial surface or the white matter surface is possible.

The MEG Core already successfully built and piloted a prototype 16-channel system, which is fixed together in a 4x4 sensor, custom 3D printed rigid array. This array can be positioned over sensory cortices and used to demonstrate localization of sensory evoked potentials. Three additional sensors serve as reference sensors to remove low frequency environmental interference.

The object of the current proposal is to expand the prototype system from 16 channels to 56 channels, in order to encompass a greater extent of the cortex. To accomplish this, we plan to purchase 45 sensors and 2 control electronics chassis.

Similar to our pilot device, sensors would be positioned in a dense array, with sensor positions fixed relative to each other. Our plan to orient our sensors immediately adjacent to each other has the obvious advantage of obtaining maximal spatial orientation. This arrangement, however, results in increased cross-talk between sensors.

In contrast to systems where individual sensors can be moved into various positions, our array will be fixed, such that the orientation of the sensors relative to each other does not change. This would enable exact measurement of the cross-talk between sensors, and subsequent calibration of the resultant data, minimizing the influence of cross-talk on measurements. This requires additional electronics which require synchronization with measurements from the sensors.

3. Generic Name of Product: Optically pumped magnetometers (OPMs), and control electronics chassis

4. Purchase Description: FieldLine, Inc. O119 PM sensors and electronics chassis or “equal” Optically pumped magnetometers (OPMs), and control electronics chassis.

4. Salient characteristics

• Must be compatible with the existing FieldLine, Inc. OPM sensors and electronics.

• The OPM sensors are zero-field sensors, using nulling fields to counteract any residual background field. In addition, sensors should operate using feedback coils to maintain zero field at the active region of the sensor to maximize linearity.

• Sensor footprint – 13 x 15mm, this determines the minimum spacing of sensors and thus contributes to the minimum spatial resolution achievable with the system.

• The sensor surface should be of a temperature that allows safe placement directly on the scalp.

• Sensor sensitivity – must be at or below 15 fT/√Hz.

• Sensor bandwidth should be at least 200Hz.

• Sensor electronics should allow individual control over the feedback coils of each sensor, to allow for calibration.

• Sensor electronics should synchronize output of the sensors and allow custom channels for ADCs and triggers.

• Acquisition software must operate in a Linux environment.

5. Quantity: Forty-five (45) optically pumped magnetometers and two (2) control electronics chassis.

6. Delivery Date: 60 days.

7. Period of Performance: Standard recommended.

8. Option(s): No options requested.