Skip to content

Adaptation of Neuropixels Probe Permits High-Resolution Neural Recording in Humans

Key findings

  • The Neuropixels digital neural probe, previously used for research in small animals, produces unprecedently clear, precise recordings of single-neuron activity
  • Researchers at Massachusetts General Hospital developed a thicker, sturdier version of the probe and made technological advancements so it could be tested in patients already scheduled for craniotomies
  • Successful recordings were made from one patient undergoing epilepsy surgery and two patients during implantation of deep brain stimulator leads (one awake and one under general anesthesia)
  • Recordings were clear and stable within minutes after the probe was inserted into cortical tissue; in two cases, the activity of more than 200 neurons was recorded over five to 10 minutes
  • With continued methodological improvements, the use of the Neuropixels probe in humans should allow increasingly detailed explorations of cortical function in both health and pathology, and accelerate the development of neurorestorative technologies

In 2017, an international collaboration of scientists and charitable foundations introduced Neuropixels, a digital neural probe designed to facilitate in vivo research in small animals. It produced unprecedently clear, precise recordings, and the tool was rapidly adopted. It has been continuously improved for application in diverse species, including non-human primates.

Now, researchers at Massachusetts General Hospital have adapted the Neuropixels system for recording single-neuron activity in humans. Technology to do that was already available, but the quality has lagged far behind what's been possible in animals.

Angelique C. Paulk, PhD, instructor in Neurology, and Sydney Cash, MD, PhD, co-director of the Center for Neurotechnology and Neurorecovery (CNTR) in the Department of Neurology at Mass General, and colleagues provide a detailed report in Nature Neuroscience. They describe the techniques they used to make three successful recordings and they describe lessons learned from unsuccessful cases.

Adapting the Probe

Neuropixels 1.0-S is a silicon array, 24 µm thick × 70 µm long × 10 mm wide, with 384 user-selectable recording channels and 960 microelectrode contacts.

The researchers attempted neural recordings in nine men and women undergoing placement of deep brain stimulator (DBS) leads for the treatment of a movement disorder or removal of brain tissue for treatment of cancer or epilepsy.

In initial experiences with two patients, the device fractured, so the team developed an improved probe, Neuropixels 1.0-ST, that's 100 µm thick (the other dimensions are the same).

Four other sets of recordings were unsuccessful due to excessive noise. Thus, another major adaptation the researchers made was to identify and reduce external sources of noise (e.g., anesthesia intravenous pumps) as well as internal sources (e.g., by separating ground and reference, unlike what's done in mouse studies).

Other technical developments for human use of the probe were:

  • Sterilizing it with ethylene oxide and maintaining sterile conditions in the operating room
  • Mounting it to a neurosurgical robot or sterile micromanipulator
  • Stereotactically guiding its insertion through a burr hole or craniotomy window
  • Using a semi-automated post hoc registration method to isolate single neurons and track them over time, since patient safety considerations preclude suppressing the brain

Three Successes

Successful recordings were made from one patient undergoing epilepsy surgery (under general anesthesia, lateral temporal lobe) and two patients during implantation of DBS leads (one awake and one under general anesthesia, dorsolateral prefrontal cortex).

Within minutes after inserting Neuropixels into cortical tissue, the researchers were able to clearly observe and stably record the activity of large numbers of neurons. In two cases, they recorded more than 200 neurons over five to 10 minutes; by comparison, older technology captures 10 to 150 neurons and typically well below 150.

Because of the very high spatiotemporal resolution, it was possible to distinguish separate neurons based on waveform shapes as well as classify the waveforms using objective techniques. The team also reports evidence the probe could be used to provide detailed microcircuit maps of the human cortex.

Looking Ahead

The next step for the Neuropixels probe will be experiments in which awake patients engage with a task during the recording period. Over time, the high-resolution, microscale, single-neuron information it provides is expected to allow for a deeper understanding of:

  • Healthy cortical function
  • The pathophysiology of neurologic diseases and brain tumors, and cancer metastasis to the brain
  • The effects of anesthetic agents
  • The therapeutic possibilities of brain–computer interfaces, including those for controlling prosthetic limbs and reanimating upper extremities
  • Cognition itself

Learn more about the Center for Neurotechnology and Neurorecovery

Learn more about research in the Department of Neurology


Jimmy C. Yang, MD, and Sydney S. Cash, MD, PhD, of the Department of Neurology, and colleagues made microelectrode recordings during neurosurgeries and identified cortex microdomains that seem to contribute to the seizure network—an important step toward understanding how to develop better treatments for epilepsy.


Neurologists at Massachusetts General Hospital have demonstrated that measuring different types of connectivity in the brain can have entirely different implications for the interpretation of brain function as well as for understanding connectivity relative to clinical diagnoses.