Non-invasive deep brain stimulation

Since 1997, more than 100,000 Parkinson’s Disease patients have been treated with deep brain stimulation (DBS), a surgical technique that involves the implantation of ultra-thin wire electrodes. The implanted device, sometimes referred to as a ‘brain pacemaker’, delivers electrical pulses to a structure called the subthalamic nucleus, located near the centre of the brain, and effectively alleviates many of the physical symptoms of the disease, such as tremor, muscle rigidity, and slowed movements.

DBS is generally safe but, like any surgical procedure, comes with some risks. First and foremost, it is highly invasive, requiring small holes to be drilled in the patient’s skull, through which the electrodes are inserted. Potential complications of this include infection, stroke, and bleeding on the brain. The electrodes, which are implanted for long periods of time, sometimes move out of place; they can also cause swelling at the implantation site; and the wire connecting them to the battery, typically placed under the skin of the chest, can erode, all of which require additional surgical procedures.

Now, researchers at the Massachusetts Institute of Technology have a developed a new method that can stimulate cells deep inside the brain non-invasively, using multiple electric fields applied from outside the organ. In a study published today in the journal Cell, they show that the method can selectively stimulate deep brain structures in live mice, without affecting the activity of cells in the overlying regions, and also that it can be easily adjusted to evoke movements by stimulation of the motor cortex.

The new method, called temporal interference, exploits the fact that neurons do not respond to electric fields with frequencies of around 1,000 Hertz (Hz, or cycles per second) or more. Thus, high frequency electric fields applied to the brain pass through it without affecting neuronal activity. If, however, two fields are applied to the brain, at high frequencies that differ by small amounts corresponding to the frequencies to which neurons can respond, they interfere with each other to produce an ‘envelope’ electric field that excites the cells within it.

For example, applying two opposing fields, with frequencies of 2000 and 2010Hz, produces an envelope field with a frequency of 10Hz wherever the two high frequency fields cross paths. This lies within the frequency range to which neurons respond, and so stimulates neurons lying beneath the envelope to fire. Nir Grossman and his colleagues at MIT’s Synthetic Neurobiology Group therefore reasoned that it might be possible to generate such low frequency electric field envelopes deep inside the brain, which would stimulate nerve cells in the envelopes without stimulating those on top, which are exposed to either one of the high frequency fields used to generate the envelope, but not both.

The researchers used computer models to simulate the effects of their technique, and then tested it in anaesthetised mice, aiming their electrodes at the hippocampus, a region lying deep within the temporal lobes, which is critical for learning and memory. They used automated patch clamping to show that stimulation activates cells in the envelope, then dissected the animals’ brains and used fluorescently-labelled antibodies to visualise the activity of c-Fos, a so-called ‘immediate-early’ gene that is switched on rapidly when neurons fire. This revealed c-Fos expression in the region of the hippocampus targeted by the electrodes, but not in other hippocampal regions, or in overlying regions of the cerebral cortex, confirming that the method specifically activates neurons in the low frequency electric field envelope generated at the intersection of the two high frequency fields (see image above).

To assess the safety of the technique, Grossman and his colleagues stimulated the hippocampus in awake mice and then stained the animals’ brains with antibodies that bind to proteins that are synthesized by dying cells. They also used thermometer probes placed immediately underneath the electrodes to measure brain temperature during the procedure. This revealed that the method did not alter the density of neurons, or the numbers of dying cells, and that the high frequency electrical fields did not increase brain tissue temperature beyond the normal range. Nor did any of the mice experience siezures – another complication of invasive DBS – during or after application of the electric fields to their brains.

Finally, the researchers made small adjustments to their stimulating electrodes, and steered them towards the primary cortex, which controls voluntary movement. Stimulation of the left motor cortex region associated with forepaw movements produced movements of the animals’ right paws, while stimulation of other regions of the motor cortex caused the ears and specific whiskers on the opposite side of the animals’ bodies to twitch.

The new technique has obvious advantages over deep brain stimulation. It also has advantages over existing non-invasive brain stimulation methods, such as transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS). “With TMS and tDCS you can activate deep regions, but you also can activate overlying ones, and that could cause unwanted side effects,” says senior author Ed Boyden. “Targets for disorders such as depression, Alzheimer’s, PTSD, and so forth, are deep in the brain, and they might be more selectively stimulatable with our method.”

Boyden adds that the method is already being tested in humans. “We’ve already begun some human stimulation trials with normal healthy volunteers, although it’s very early days and very exploratory,” he says, “and we are now reaching out to experts on epilepsy, tinnitus, depression, and other disorders to see if we can help.”


Grossman, N., et al. (2017). Noninvasive Deep Brain Stimulation via Temporally Interfering Electric Fields. Cell169: 1029-41. [Full text]

Header image: Temporal interference stimulation excites an area in the mouse hippocampus, shown in bright green through c-Fos labelling. (Nir Grossman, Suhasa B. Kodandaramaiah and Andrii Rudenko)