Using External Electrodes to Treat Tremors

The most common cause of tremor is not Parkinson's disease. Rather, essential tremors is much more common, with 4% of the population over 40 years of age suffering. As one ages, Essential tremor only becomes more likely, with 1 in 5 people over the age of 90 afflicted. Given that the disease is so common, one would expect that their is a easy medical intervention. Sadly, that is not the case. There seems to be a diverse set of causes for Essential Tremor, and as such many medicines just don't work on a significant subset of patients. As such it is critical to develop new treatments to treat the disease.
Fig.1: Example of a patient with Essential tremor (left) trying to draw a spiral.

A new study from Dr. Sebastian Schreglmann and Colleges does just that. Using the novel and non-invasive transcranial electrical stimulation, they were able to suppress tremors in half of patients. This method uses external electrodes attached to the base of the head  (much like an EEG) to  create an external electrical field. This electrical field disrupts the firing of neurons near the surface of the cerebellum. Luckily, essential tremor is caused by a pathological synchronization of Purkinje cells in the cerebellum. Fortunately for us, the Purkinjie cells lie on the surface of the cerebellum, and are easy to stimulate and disrupt. 
Fig.2: Example of the treatment set up. By recording the phase of the tremor in figure a, the electrode (placed over the cerebellum) can inject a sinusoidal AC current. This alleviates the tremor.

However, one cannot slap an electrode at the base of the head, and call it a day. The phase of the electrical current must be aligned to the phase of the current. That is, the electrode emits a sinusoidal AC current, whose phase is aligned to the phase of the patients tremor.  Determining the phase of the tremor is no easy matter. The researchers developed a novel algorithm, the endpoint-corrected Hilbert transform, that allows for real time computation of the phase of an oscillator (in this case a accelerometer attached to the patients wrist).

A normal Hilbert transform can be used to compute the phase of an oscillator.  However, this cannot run in real time. Algorithms that can run in real time often suffer from a cutoff of the waveform (as one cannot know the waveform for future times). This results in a sudden discontinuous jump, and for those familiar with Fourier series, the Gibbs phenomena creates a large error at the discontinuity. To get around this, the authors use there endpoint correct Hilbert transform, that can remove the discontinuity and avoid the Gibbs phenomena. 

Fig.3: An example of the signal computed with the Hilbert Transform (b) and its computed phase error (c). This is contrasted with the much more accurate endpoint-corrected Hilbert Transform(ecHT) (d) as evident by its lower error (e)

Once they compute the phase of the tremor, they simply adjust the transcrinal current to have the same frequency and phase delay as the tremor. Once everything is dialed in, the tremor amplitude is significantly depressed. Happily for the patient, they can now move without being crippled by a tremor. 

As mentioned this only works for about half of the patients. It is a bit of an open question why only half of the patients respond positively. However, it seems that patients where the treatment work all have a few similarities in there tremors. Using machine learning techniques, they were able to classify which patients would respond, and which would not respond. To oversimplify a bit, it seems to boil down to which phase is easy to predict. Paradoxically, easier predictions are harder to suppress, potentially implying higher pathological synchronization in the cerebellum.

Nonetheless, for those patients who do respond, the method works very well. This offers a new, non-medication based alternative to treatment.


Author: Alexander J. White


Source: Schreglmann, S.R., Wang, D., Peach, R.L. et al. Non-invasive suppression of essential tremor via phase-locked disruption of its temporal coherence. Nat Commun 12, 363 (2021). https://doi.org/10.1038/s41467-020-20581-7

留言