Probe reliably records activity of large neuron populations in brains of non-human primates

To map the mammalian brain and its various functions with increasing precision, neuroscientists rely on high-resolution imaging techniques and other advanced experimental tools. These include high-density silicon probes, needle-like devices integrating several electrodes that can be inserted into brain tissue to pick up voltage changes associated with the firing of neurons.

These devices have so far mostly been used to monitor and study the activity of neurons in the rodent brain. However, they proved less effective when studying the brains of non-human primates (NHPs), such as macaques, which more closely resemble the human brain.

Researchers at Columbia University Medical Center and Columbia University recently demonstrated the potential of Neuropixels 1.0 NHP, a more scalable probe developed by IMEC, for collecting brain-wide and high-resolution neural recordings in macaques and other NHPs.

Their project resulted in a paper published in Nature Neuroscience.

“Our paper was inspired by a fundamental need in the neuroscience community to overcome the limitations of existing neural recording technologies in NHPs,” Eric M. Trautmann, first author of the paper, told Medical Xpress.

“The original Neuropixels 1.0 probe, while transformative for rodent models, had limited functionality in NHPs due to its 10-mm length, which restricted access to superficial brain targets, and a thin shank (24 µm) that made insertion through the primate dura mater difficult. Other existing linear arrays are limited in channel count, and surface arrays like the Utah array or floating microwire arrays were confined to superficial cortex and fixed depths.”

Trautmann and his colleagues were looking for a probe that would allow them to accurately and reliably record neural activity in the brains of NHPs and other large animals. They wanted this device to also reliably record the activity of large populations of neurons with single-neuron and single-spike resolution, or in other words, isolate the activity of single neurons and detect individual action potentials.

“The new recording electrode developed by IMEC is fabricated as a 54-mm-long monolithic piece of silicon that integrates both the shank and the base electronics,” explained Trautmann.

“The probe features a significantly longer, wider, and thicker shank (45 mm long, 125 µm wide, and 90 µm thick) compared to the original Neuropixels 1.0 probe.”

The probe tested by the authors has a total of 4,416 recording sites (i.e., pixels) distributed along the full length of its 45-mm shank and grouped into 11.5 “banks” of 384 channels each. Notably, experimenters can also select one of these “banks” to collect simultaneous recordings across 384 sites, via a switch located under every site.

“The probe utilizes the same low-noise readout channels with programmable gain and 10-bit resolution on a 130-nm Silicon-on-Insulator CMOS platform as the Neuropixels 1.0,” said Trautmann.

“The tip of the shank is mechanically ground to a 25° bevel angle on the side plane, creating a sharpened tip along both axes to facilitate insertion and minimize tissue damage.”

The new Neuropixels device offers neuroscientists deeper access into structures throughout the brain of large animals, which cannot be accessed by shorter high-density probes. Moreover, its 90 µm thickness and sharper shank allow it to effectively penetrate the tough membrane covering the primate brain and spinal cord, known as the primate dura mater.

“Further advantages of our approach include its scalability and high channel count, or in other words, the ability to programmably select 384 channels from over 4,400 sites and collect simultaneous multi-area recording from thousands of neurons,” said Trautmann.

“This feature allows experimenters to decouple the process of optimizing a recording location from probe positioning, facilitating surveys of neural activity along the entire probe length without physical movement.”

As the newly developed probe is larger than the individual reticules (i.e., masks) commonly used to fabricate devices via photolithographic techniques, its developers had to devise approaches to enable its reliable realization. These included a “stitching” strategy to precisely align features along the probe across multiple steps, as well as stress compensation layers that prevent it from bending.

“Internal metal wires were widened and spaced to minimize resistance, thermal noise, and crosstalk,” explained Trautmann. “Another advantage of our probe is its high spatial resolution. The dense recording sites facilitate high-quality, automated spike sorting and enable continuous tracking of neurons even with drifting motion between the probe and tissue.”

Compared to existing technologies, the Neuropixels 1.0 NHP probe is also more cost-effective. It could thus dramatically lower the cost associated with the recording of neurons, with a total system cost ranging between US$7,000 and US$15,000.

“We demonstrated the probe’s potential through four example experiments in macaques, addressing diverse neuroscience questions,” said Traumann.

“We showed large-scale surveys of retinotopic organization across multiple extrastriate visual cortical areas, recording thousands of neurons simultaneously and illustrating the orderly shift of receptive fields across cortical depths.

“We also collected stable, large-scale recordings during motor behaviors in superficial and deep structures (e.g., primary motor cortex, premotor cortex, globus pallidus interna, supplementary motor area), demonstrating the capture of diverse temporal patterns and improved force prediction with more neurons.”

Using the new probe, the researchers could also record activity in an area of the NHP brain that is typically difficult to access (i.e, deep inferotemporal cortex face patches), all while the animals were looking at images of faces. During a single data collection session, they were able to detect hundreds of neurons that appeared to contribute to the recognition of faces, which would have previously taken years.

“We also demonstrated the probe’s utility for studying single-trial correlates of decision-making, showing how LIP and superior colliculus (SC) neuron populations track accumulated evidence and exhibit distinct dynamics only observable through high-yield single-trial analyses,” said Trautmann.

“The high density also facilitated measuring spike-spike correlations between neuron pairs, which is indicative of synaptic connections or shared input, enabling mapping of putative connections across cortical laminae and between regions.”

This recent study could soon open new valuable possibilities for neuroscience research involving large animals. By overcoming various widely reported engineering challenges, the new probe developed at IMEC was found to reliably record the activity of neurons across the brains of non-human primates.

“We successfully surmounted significant engineering challenges related to fabricating large-scale silicon probes that exceed standard photolithographic reticle sizes through methods like stitching and by designing for mechanical strength and flexibility (stress compensation layers) for the long shank,” said Trautmann.

“This technology enables new classes of experiments previously deemed impractical or impossible. This includes detailed electrophysiological mapping of brain areas at single-neuron and single-spike resolution, measuring spike-spike correlations between cells, and conducting simultaneous brain-wide recordings at an unprecedented scale.”

In the future, the Neuropixels 1.0 NHP probe could be deployed in other laboratories worldwide, reducing the efforts, resources and costs currently associated with studying the brains of large animals. It could also enable more advanced and complex neuroscience experiments involving NHP, which could lead to new exciting discoveries.

“Our probe facilitates a more comprehensive and unbiased mapping of neural activity across multiple brain regions and depths, which is essential for understanding the coordinated action of large neuronal populations involved in sensory, motor, and cognitive operations,” said Traumann.

“It could also allow researchers to perform single-trial analyses of neural activity at high resolution, which is particularly critical for studying cognitive functions, where brain processes can vary across task repetitions.”

If used to map neural circuits, the Neuropixels 1.0 NHP probe could soon also help to better understand how anatomical circuits in the brains of large mammals perform specific computations. While the current version of the device is primarily optimized to perform acute recordings, IMEC is currently working on further improving its capabilities.

“As part of our next studies, we plan to develop new hardware and test semi-chronic (multiple days or weeks) implantation,” added Trautmann.

“This will require new implant designs and is currently an untested but conceivable capability with appropriate insertion methods and hardware. Moreover, we are working on developing probes with intracortical microstimulation (ICMS) capabilities, which is not yet possible. However, the current generation of probes can be used to record alongside stimulation via a separate electrode.”

© 2025 Science X Network

To map the mammalian brain and its various functions with increasing precision, neuroscientists rely on high-resolution imaging techniques and other advanced experimental tools. These include high-density silicon probes, needle-like devices integrating several electrodes that can be inserted into brain tissue to pick up voltage changes associated with the firing of neurons.

These devices have so far mostly been used to monitor and study the activity of neurons in the rodent brain. However, they proved less effective when studying the brains of non-human primates (NHPs), such as macaques, which more closely resemble the human brain.

Researchers at Columbia University Medical Center and Columbia University recently demonstrated the potential of Neuropixels 1.0 NHP, a more scalable probe developed by IMEC, for collecting brain-wide and high-resolution neural recordings in macaques and other NHPs.

Their project resulted in a paper published in Nature Neuroscience.

“Our paper was inspired by a fundamental need in the neuroscience community to overcome the limitations of existing neural recording technologies in NHPs,” Eric M. Trautmann, first author of the paper, told Medical Xpress.

“The original Neuropixels 1.0 probe, while transformative for rodent models, had limited functionality in NHPs due to its 10-mm length, which restricted access to superficial brain targets, and a thin shank (24 µm) that made insertion through the primate dura mater difficult. Other existing linear arrays are limited in channel count, and surface arrays like the Utah array or floating microwire arrays were confined to superficial cortex and fixed depths.”

Trautmann and his colleagues were looking for a probe that would allow them to accurately and reliably record neural activity in the brains of NHPs and other large animals. They wanted this device to also reliably record the activity of large populations of neurons with single-neuron and single-spike resolution, or in other words, isolate the activity of single neurons and detect individual action potentials.

“The new recording electrode developed by IMEC is fabricated as a 54-mm-long monolithic piece of silicon that integrates both the shank and the base electronics,” explained Trautmann.

“The probe features a significantly longer, wider, and thicker shank (45 mm long, 125 µm wide, and 90 µm thick) compared to the original Neuropixels 1.0 probe.”

The probe tested by the authors has a total of 4,416 recording sites (i.e., pixels) distributed along the full length of its 45-mm shank and grouped into 11.5 “banks” of 384 channels each. Notably, experimenters can also select one of these “banks” to collect simultaneous recordings across 384 sites, via a switch located under every site.

“The probe utilizes the same low-noise readout channels with programmable gain and 10-bit resolution on a 130-nm Silicon-on-Insulator CMOS platform as the Neuropixels 1.0,” said Trautmann.

“The tip of the shank is mechanically ground to a 25° bevel angle on the side plane, creating a sharpened tip along both axes to facilitate insertion and minimize tissue damage.”

The new Neuropixels device offers neuroscientists deeper access into structures throughout the brain of large animals, which cannot be accessed by shorter high-density probes. Moreover, its 90 µm thickness and sharper shank allow it to effectively penetrate the tough membrane covering the primate brain and spinal cord, known as the primate dura mater.

“Further advantages of our approach include its scalability and high channel count, or in other words, the ability to programmably select 384 channels from over 4,400 sites and collect simultaneous multi-area recording from thousands of neurons,” said Trautmann.

“This feature allows experimenters to decouple the process of optimizing a recording location from probe positioning, facilitating surveys of neural activity along the entire probe length without physical movement.”

As the newly developed probe is larger than the individual reticules (i.e., masks) commonly used to fabricate devices via photolithographic techniques, its developers had to devise approaches to enable its reliable realization. These included a “stitching” strategy to precisely align features along the probe across multiple steps, as well as stress compensation layers that prevent it from bending.

“Internal metal wires were widened and spaced to minimize resistance, thermal noise, and crosstalk,” explained Trautmann. “Another advantage of our probe is its high spatial resolution. The dense recording sites facilitate high-quality, automated spike sorting and enable continuous tracking of neurons even with drifting motion between the probe and tissue.”

Compared to existing technologies, the Neuropixels 1.0 NHP probe is also more cost-effective. It could thus dramatically lower the cost associated with the recording of neurons, with a total system cost ranging between US$7,000 and US$15,000.

“We demonstrated the probe’s potential through four example experiments in macaques, addressing diverse neuroscience questions,” said Traumann.

“We showed large-scale surveys of retinotopic organization across multiple extrastriate visual cortical areas, recording thousands of neurons simultaneously and illustrating the orderly shift of receptive fields across cortical depths.

“We also collected stable, large-scale recordings during motor behaviors in superficial and deep structures (e.g., primary motor cortex, premotor cortex, globus pallidus interna, supplementary motor area), demonstrating the capture of diverse temporal patterns and improved force prediction with more neurons.”

Using the new probe, the researchers could also record activity in an area of the NHP brain that is typically difficult to access (i.e, deep inferotemporal cortex face patches), all while the animals were looking at images of faces. During a single data collection session, they were able to detect hundreds of neurons that appeared to contribute to the recognition of faces, which would have previously taken years.

“We also demonstrated the probe’s utility for studying single-trial correlates of decision-making, showing how LIP and superior colliculus (SC) neuron populations track accumulated evidence and exhibit distinct dynamics only observable through high-yield single-trial analyses,” said Trautmann.

“The high density also facilitated measuring spike-spike correlations between neuron pairs, which is indicative of synaptic connections or shared input, enabling mapping of putative connections across cortical laminae and between regions.”

This recent study could soon open new valuable possibilities for neuroscience research involving large animals. By overcoming various widely reported engineering challenges, the new probe developed at IMEC was found to reliably record the activity of neurons across the brains of non-human primates.

“We successfully surmounted significant engineering challenges related to fabricating large-scale silicon probes that exceed standard photolithographic reticle sizes through methods like stitching and by designing for mechanical strength and flexibility (stress compensation layers) for the long shank,” said Trautmann.

“This technology enables new classes of experiments previously deemed impractical or impossible. This includes detailed electrophysiological mapping of brain areas at single-neuron and single-spike resolution, measuring spike-spike correlations between cells, and conducting simultaneous brain-wide recordings at an unprecedented scale.”

In the future, the Neuropixels 1.0 NHP probe could be deployed in other laboratories worldwide, reducing the efforts, resources and costs currently associated with studying the brains of large animals. It could also enable more advanced and complex neuroscience experiments involving NHP, which could lead to new exciting discoveries.

“Our probe facilitates a more comprehensive and unbiased mapping of neural activity across multiple brain regions and depths, which is essential for understanding the coordinated action of large neuronal populations involved in sensory, motor, and cognitive operations,” said Traumann.

“It could also allow researchers to perform single-trial analyses of neural activity at high resolution, which is particularly critical for studying cognitive functions, where brain processes can vary across task repetitions.”

If used to map neural circuits, the Neuropixels 1.0 NHP probe could soon also help to better understand how anatomical circuits in the brains of large mammals perform specific computations. While the current version of the device is primarily optimized to perform acute recordings, IMEC is currently working on further improving its capabilities.

“As part of our next studies, we plan to develop new hardware and test semi-chronic (multiple days or weeks) implantation,” added Trautmann.

“This will require new implant designs and is currently an untested but conceivable capability with appropriate insertion methods and hardware. Moreover, we are working on developing probes with intracortical microstimulation (ICMS) capabilities, which is not yet possible. However, the current generation of probes can be used to record alongside stimulation via a separate electrode.”

© 2025 Science X Network