February 6, 2015
NeuroNexus has developed a true 3-dimensional probe, the Matrix Array. The Matrix Array is a silicon-based probe that has a 2-dimensional array of shanks in the X-Y plane, and each shank has multiple recording sites aligned vertically on a Z-axis. The Matrix Array, then, allows for recording volumes of tissues that include multiple neural layers as well as multiple columns.
Figure 1: (Left) Render of 3D nature of the Matrix Array. Multiple 2D probes (in this example, spanning 1400 µm along z-axis and 1200 µm on x-axis) are stacked to span 1800 µm in the y-axis. The stacking spacing can be adjusted to span 600 µm or 3000 µm in the y-axis, and different 2D probes can be selected to span different recording areas in the x-axis and z-axis. (Right) Photograph of a Matrix Array. The small white circles along each shank are the recording sites.
These blog posts describe some of the developmental process of the Matrix Array, and the testing that we have done to ensure that it is a high quality product. Matrix Arrays have been field tested in various areas of the cortex of non-human primates at five independent research institutions. In our last blog entry we reported some of the developments in our Matrix Array insertion procedures that were tested and re-designed based upon work done in labs in Michigan and Texas. Today, we’ll focus on the progress of the recording tests that have been done to date.
In the spring of 2014 a lab in Illinois implanted a 128-channel Matrix Array into the primary motor cortex. The array was comprised of four 32-channel (M4x8-2mm-200-400-703: four shanks, eight sites per shank, 2 mm long shanks, 200 µm site spacing, 400 µm shank spacing, 703 µm2 site area) arrays spaced 1000 µm apart. The experiment lasted for two months, during which they recorded spontaneous activity during periods of rest and examined the power spectral density. This experiment was reported on a poster at the Neural Interfaces Conference in June, 2014.
Figure 2: Neural Recordings from the primary motor cortex of a non-human primate. (Left) Sample of recorded waveforms from a bank of 32 electrode contacts (t = 28 days post-implant). (Right) The number of tracked single units from the same bank of 32 electrode contacts over the first month.
Two labs are currently recording from Matrix Array Implants performed in October of 2014. One lab, in Maryland, implanted a 128-channel Matrix Array comprised of four 32-channel M4x8-2mm-200-400-703 probes, spaced 1000 µm apart. They are recording both LFP and single unit data.
A third lab, also in Illinois, implanted two 128-channel Matrix Arrays (256 channels total). Both arrays had one 32-channel array that was longer (M4x8-5mm-150-200-703, 5 mm long with 150 µm site spacing, 200 µm shank spacing, site size 703 µm2) for recording in a sulcus. The other three 32-channel arrays that comprised each Matrix Array were M4x8-2mm-200-400-703 arrays. The 32-channel arrays were spaced 600 µm apart.
Figure 3: Matrix Array implanted into Area 2 and Area 3a of cortex, with cortical landmarks, arrays, and array banks labeled (anterior is to the right of the image).
At the time of this blog, both experiments have surpassed the 3-month mark. In the words of the second lab from Illinois, “Our implants have been consistently picking up neurons for the past three months, in both Area 2 and 3a. The arrays seem to have lasted much longer than other similar multi-contact electrode arrays we've had experience with, including LMAs and Michigan probes, so the Matrix array is promising for a few experiments we've wanted to conduct in cortical areas that lie deep in a sulcus, like the Area 3a experiments we've been collecting data for with this implant.”
Figure 4: Composite image of recordings taken from each bank of the two 128-channel Matrix Arrays (256 channels total) implanted into Area 2 and Area 3a, labeled in the same way as in Figure 3.
September 15, 2014
NeuroNexus has developed a true 3-dimensional probe, the Matrix Array. The Matrix Array is a silicon-based probe assembly that is comprised of an array of shanks aligned along a single lateral plane, with each shank containing multiple electrodes vertically positioned. The Matrix Array, then, allows for recording volumes that span both cortical layers and cortical columns. These blog posts will describe some of the developmental process of the Matrix Array, and the testing that we have done to ensure that it is a high quality product. Today, we’ll focus on the methods for implanting the Matrix Array.
Traditional NeuroNexus probes have either a single shank or multiple shanks that are all positioned in the same plane. Thus, insertion of those probes is fairly straight forward. However, probes with a 2-dimensional array of shanks come with special insertion challenges. Depending on how close the shanks are to each other, it is possible to get a “pincushion effect,” in which the penetration of one shank is hindered by tissue dimpling caused by a different shank. (This happened to me with a probe that I was attempting to use during my first postdoc to record in the dorsal root ganglia, and the shanks of the probe bent instead of inserting, ruining the probe.) One way to overcome this effect is to insert the probe with a lot of force, and this is a method that is often used to insert probes with this shape profile. However, high-force insertion can cause tissue damage that manifests in longer healing times, longer periods before recordings can be taken, or possible neural damage. Thus, we spent some time developing and testing alternate insertion methods for the Matrix Array.
The shanks of the Matrix array are thicker (50 µm) than the standard NeuroNexus probe (15 µm). This increased thickness makes the shanks stronger and able to cleanly penetrate tissue. The shanks are still thin enough, however, that the ratio of space between shanks to shank thickness still prevents the pincushion effect. Thus, we are able to do controlled insertion of the Matrix Array without requiring a great deal of force. Finally, we chose a computer-controlled insertion motor with extremely fine step resolution (0.05 µm/step) and speed resolution (0.22 µm/sec).
We first tested our insertion method on models, such as plastic wrap over agar. Once thoroughly tested in that way, we moved onto in vivo testing and eventually implantations with our beta testers.
One such test was done in the lab of researchers in Texas in October of 2013 (pictured above). The surgery was performed by our collaborators, implanting the Matrix Array into the motor cortex of a rhesus macaque monkey. The Matrix Array is held onto the tip of the IST-Motor insertion tool by vacuum. Using a craniotomy 1.6 cm in diameter, we fully inserted the 0.75 mm probe, in increments of 0.2 mm every 30 s. Once satisfied that the probe was completely inserted we turned off the vacuum and withdrew the insertion tool, leaving the probe in the brain. During this experiment we held forceps on the back of the inserted Matrix Array to make sure that it remained in place, but in prior and subsequent tests the forceps were not necessary. At this point, the probe was successfully inserted.
The insertion process is but one step on the path to having successful 3-dimensional recordings, and it perhaps is not even the first step that one might think of. But it is an important step nonetheless, and we made sure to develop a good approach and fully test it to ensure that the Matrix Array provides maximum benefit with minimal energy.
August 12, 2014
Optogenetics has been one of the most exciting fields in neuroscience with tremendous efforts undertaken to advance the techniques and tools. NeuroNexus adapted early to provide researchers with innovative optogenetics products, exemplified by the industry leading Optoelectrode product line. NeuroNexus Optoelectrodes have been continuously evolving products, starting from the very first Optoelectrodes that featured basic LC connectors with limited packaging options. Since then we have used extensive design insight as well as refinement of optics to fine-tune and further develop the Optoelectrode, resulting in the NeuroNexus Coupler (NNc) and supporting accessories such as patch-cords and implantable fibers. The NeuroNexus Optoelectrode has become the standard research tool for simultaneous optical stimulation and neural recordings. We continue to research new areas to provide critical value to our Optogenetics customers and are excited to share the addition of two new in-house capabilities: (1) introduction of etched smaller diameter fibers and (2) robust, high-fiber count packaging.
Left: Dual-fiber Optrode; Middle: Quad-fiber Optrode, with lasers attached;
Right: Close-up of Quad-fiber Optrode microelectrode array
Etched, Smaller Diameter Fibers
NeuroNexus has historically provided Optoelectrodes with our standard fiber with the following specification: 105 µm core, 125 µm cladding, 0.22 NA. We have since then introduced the 0.66 NA fiber to provide a solution to customers requiring higher numerical aperture. Our next development project for our fibers was to provide a smaller diameter fiber Optoelectrode. Smaller diameter fibers would theoretically attach better to our 15 µm thin microelectrode arrays as the flexibility of smaller diameter fibers provide better mechanical matching with the narrow shanks. We also expected insertion to improve with less dimpling. Lastly, tissue damage would decrease significantly as the cross-sectional area of the fiber would decrease.
With input from Dr. Kenji Mizuseki from Allen Institute for Brain Science, we pursued the approach of HF (hydrogen fluoride) etching the 50 µm core, 125 µm cladding fiber to reduce the cladding diameter. By immersing the fiber for a calculated period, we were able to successfully predict and accomplish desired cladding diameter resulting in 50 µm core, ~65 µm cladding fibers. This decrease of cladding diameter leads to a reduction in cross-sectional area of approximately 73%! We are very excited that our approach yields repeatable results that provide improved tissue interface for our Optoelectrodes.
Above: HF-etched optical fiber, shown on a SEM.
The OD has been reduced to ≈ 62.5 µm, with the core remaining at 50 µm.
Robust, High-Count Fiber Optoelectrodes
Optoelectrodes featuring multiple fibers have been popularly requested by our customers, with one-fiber-per-shank mounting on the Buzsaki design microelectrode arrays being of particular interest. While lab-built, multi-fiber optrodes have been used in several research studies, combining many fiber terminations into a minimal, robust package has been a large hurdle for commercial multi-fiber Opeoelctrode development. Our standard fibers pose additional challenges:
- Size and mechanical stiffness of our standard fibers pose assembly difficulties
- The increase in cross-sectional area with each additional fiber leads to significant tissue damage and higher risk of insertion failure
With the development of our etched fibers, the two above concerns are significantly dissipated. We were also able to solve the packaging problem by utilizing our new, precise 3D printer to design and fabricate connectors that can house many ferrule terminations without adding significant bulk to the probe package. Thus yielding a robust probe package that allows many fibers to be mounted on one Optoelectrode.
January 7, 2015
NeuroNexus recently collaborated with Shane Heiney of Javier Medina's lab at the University of Pennsylvania to design a custom component for their research with mice. Through the use of our in-house high-resolution 3D printing, we were able to arrive at a solution that improved the process and duration of Shane's experiments.
Tell us about your application and the problem your 3D print design was trying to solve.
"In our lab we make repeated daily acute microelectrode recordings from awake mice and needed a way to keep the craniotomy healthy between sessions. We had tried to chronically implant several different custom designed plastic rings to serve as 'recording chambers' but none of these designs offered a good way to completely enclose the craniotomy to keep air and infectious particles out. This forced us to perform dura peels almost daily to remove scar tissue and other 'gunk' that had accumulated between sessions.
What we needed was a chamber with a well-fitted locking cap."
Why did you choose 3D printing to produce your components?
"Recording chambers with locking caps are standard in non-human primate work and several vendors provide them, but we could find no vendors offering a similar design at a mouse scale. We thought this was a perfect use case for 3D printing."
How was your experience working with NeuroNexus?
"We designed a custom recording chamber with interlocking lid and approached NeuroNexus to see if they would be able to print it for us. Our primary concern was that some of the features would be too small even for their high resolution printer to handle. Their engineers worked with us through several iterations of the design to come up with a finished product that is far more robust than our original design. They even suggested a new feature that turned out to be critical for holding the cap in place while the mouse is in its home cage between experiments."
Tell us about the results of the collaboration.
"We have been very pleased with the results of this collaboration and the newly designed chambers have greatly increased the health of the dura between sessions. We've found that by adding a small amount of silicone elastomer to the dura between experiments, which the cap prevents from falling out, we have eliminated the need for daily dura peels. Our implants routinely last for several months."
July 30, 2015
The Vector Array™ is an all-new neural probe design optimized for large animal deep brain applications (e.g. non-human primate, porcine, etc.).
Already being successfully used in labs, the Vector Array™ can be customized for different applications, taking advantage of the versatility of NeuroNexus' silicon microelectrode technology. Here are some first-hand accounts from three research groups on the work they have been doing with the Vector Array™.
Macaque Frontal Eye Field (FEF)
Macaque MRI co-registered with grid, showing sampling across depth with LFP and spiking data
A group in Tennessee is using a 32-channel Vector Array™ with 100 µm site spacing to record from the Frontal Eye Field (FEF) of the prefrontal cortex of macaque monkeys during visual tasks. According to a member of the lab, "the Vector probe overcomes previous limitations and allows us to sample a spatial 'slice' of FEF with each penetration, which we co-register to MRI images that are aligned with our chamber grids, allowing us to map the spatial distribution of responses and observe how response properties change across locations and with respect to cortical layers."
"The figure shows the MRI for one of our monkeys (bonnet macaque) with grid co-registered. With the MRI map in place, we can select a location, and sample across depth while recording LFP and spiking data, which are plotted to the right of the MRI images. These plots show 1) the spatial distribution of LFP plots and 2) the spatial distributions of different cell types across depths/contacts (indicated by V, VM, M, classified by spiking characteristics), aligned on a schematic cortical layer representation of FEF. With this setup, we can go back to specific locations across days (indicated by Day 1 and Day 2 in response maps), and begin to characterize the spatial layout of FEF responses, as well as characterizing the source of the spatially distributed LFP signal. The inset at the top right of the figure shows the quality of sorts we are able to obtain using the Vector Array, which is critical for our ability to accurately characterize the response properties of the various cell types."
Porcine Hippocampal Laminar Fields
Left: Custom-designed 32 channel Vector Array, with seven tetrode clusters. Right: 10 second signal sample corresponding to the recording location on Vector Array.
A second group in Pennsylvania is using a custom-designed 32-channel Vector Array™ with 28 of the sites clustered into seven tetrodes to record from the hippocampus of a porcine model. According to a member of the lab, "The electrodes have been working very well for field recordings with beautiful noise free recordings from the awake behaving pig in a contained room (11X11) behavioral space."
Primate Cortex Laminar Recordings
Left: Poly2 Site layout at the tip of a 32 channel Vector Array. Right: Single units recorded by the Vector Array, corresponding to recording site location.
A group in Texas has been using a 32-channel Vector Array™ with sites arranged in a Poly2 formation to do laminar recordings from both the cortex (V1), as well as deeper structures. According to a member of the lab, "We've been working with the Polytrode Vector probes for a few months now and are very enthusiastic about the results! We can easily see your depth probes becoming the standard in non-chronic primate electrophysiology recordings over all the other available probes, especially as the field moves towards more sensitive ways of probing neural circuits by recording from all layers simultaneously, for example, which your probes would allow. It will be a great way to extend the value and use of electrophysiological recordings in a time of significant advancements in optical recording as well, by creating an advantage that optical imaging does not yet allow - complete laminar recordings."
Close-up detail, Insertion methods, and more
Learn how to effectively use the Vector Array on our YouTube channel.
Learn about all the options available to you when ordering a NeuroNexus Optrode.
A true 3D neural interface, for both acute and chronic experiments.
Record up to 256 channels, simultaneously.